AU2020407591A1 - Manabodies targeting tumor antigens and methods of using - Google Patents

Abstract

This document provides methods and materials for assessing a mammal having or suspected of having cancer and/or for treating a mammal having cancer. For example, molecules including one or more antigen-binding domains (e.g., a single-chain variable fragment (scFv)) that can bind to a modified peptide (e.g., a tumor antigen), as well as method for using such molecules, are provided.

Description

MANAbodies TARGETING TUMOR ANTIGENS AND METHODS OF USING

CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Patent Application Serial No. 62/949,220, filed on December 17, 2019 and U.S. Patent Application Serial No. 63/059,638, filed on July 31, 2020. The disclosure of the prior applications are considered part of (and are incorporated by reference in) the disclosure of this application.

STATEMENT REGARDING FEDERAL FUNDING This invention was made with government support under grant CA062924 awarded by the National Institutes of Health. The government has certain rights in the invention.

BACKGROUND

1. Technical Field

This document relates to methods and materials for assessing a mammal having or suspected of having cancer and/or for treating a mammal having cancer. For example, this document provides methods and materials for using a molecule including one or more antigen-binding domains ( e.g ., a single-chain variable fragment (scFv)) that can bind to a modified peptide (e.g., a tumor antigen) to treat a mammal having a cancer.

2. Background Information

Somatic mutations in cancer are ideal targets for cancer therapy as they are expressed only in tumor cells and not in normal cells. Targeting driver gene proteins (broadly subdivided into oncogene proteins and tumor suppressor proteins) have added benefits. First, these mutations typically occur early during the development of the tumor, thus essentially all daughter cancer cells will contain the mutation. Second, the tumor's dependence on their oncogenic-endowing capacity makes resistance less likely. Finally, driver gene proteins tend to have hotspot mutations shared among many patients, thus a therapy targeting a single mutation could be applied to a broad patient population.

Most mutant proteins, including most mutant driver gene proteins, are intracellular. While small molecules can target intracellular proteins, developing small molecules that can specifically inhibit the activity of a mutant driver gene and not its wild-type (WT) counterpart has remained out of reach for the majority of such driver gene proteins. Antibodies, which can have the capacity to distinguish a single amino acid mutation, can typically only target extracellular epitopes.

The immune system samples the intracellular contents of cells through antigen processing and presentation. Following protein proteolysis, a fraction of the resulting peptides are loaded onto a human leukocyte antigen (HLA) and sent to the cell surface where they serve as a way for T cells, via their T cell receptor (TCR), to distinguish self from non self peptides. For example, a virally-infected cell will present viral peptides in its HLA, triggering T cells to kill that cell. Similarly, in cancer, mutant peptides can be presented in an HLA on the cancer cell surface, referred to as MANAs, for Mutation- Associated Neo- Antigens. In some cases, and to varying degrees, patients may mount an anti-cancer T cell response against these mutant-peptide-HLA neoantigens, and checkpoint blockade antibodies can further augment this response. However, many patients, particularly those with a low mutational burden, cannot mount a sufficient anti-cancer T cell response. A therapy or diagnostic specifically targeting MANAs could therefore provide a truly tumor-specific method to diagnose or treat cancer.

HLA class I proteins are present on all nucleated cells. There are three classical HLA class I genes, A, B, and C, each of which are highly polymorphic. Each HLA allele has a particular peptide-binding motif, and as a result, only certain peptides will bind to certain HLA alleles.

There is a continuing need in the art to develop new methods to diagnose, monitor, and effectively treat cancers.

SUMMARY

Identification of therapeutic targets highly specific to cancer cells is one of the greatest challenges for developing an effective cancer therapy.

This document provides methods and materials for treating a mammal having cancer. For example, this document provides methods and materials for using one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide (e.g., a modified peptide present in a peptide-HLA- beta-2 microglobulin (b2M or b2M) complex) to treat a mammal having a cancer (e.g, a cancer expressing the modified peptide). In some cases, one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide ( e.g ., a modified peptide present in a peptide-HLA-P2M complex) can be administered to a mammal having a cancer (e.g., a cancer expressing the modified peptide) to treat the mammal.

As demonstrated herein, scFvs were identified that target (e.g, bind to) numerous MANAs present in HLA-restricted MANAs derived from common cancer driver mutations, including RAS Q61H/L/R and p53 R175H. Also as demonstrated herein, the scFvs were used to design bispecific antibodies capable of inducing MANA-dependent T cell activation that can lead to recognition and killing of cells (e.g, cancer cells) expressing MANAs.

MANAs can be used as highly specific cancer targets because they are not present in normal tissue(s). The ability to specifically target MANAs provides a tumor-specific method to diagnose and/or treat cancer. For example, scFvs specifically targeting MANAs can be used in full-length antibodies or fragments thereof, antibody drug conjugates (ADCs), antibody radionuclide conjugates, T cells expressing a chimeric antigen receptor (CARTs), or bispecific antibodies to diagnose and/or treat a mammal having cancer. Further, an antibody that can bind to a MANA (a MANAbody), or a fragment thereof capable of binding to a MANA, have the potential of becoming widely applicable and genetically predictable off- the-shelf targeted cancer immunotherapy.

In general, one aspect of this document features molecules having an antigen-binding domain that can bind to a peptide-HLA-P2M complex, where the peptide can be derived from a modified p53 polypeptide. The modified p53 polypeptide can include from 7 amino acids to 25 amino acids (e.g, the modified p53 polypeptide can include 9 amino acids). The modified p53 polypeptide can include the amino acid sequence set forth in SEQ ID NO:l.

The antigen binding domain can include an amino acid sequence set forth in any one of SEQ ID NO: 137, SEQ ID NO: 138, SEQ ID NO: 139, SEQ ID NO: 140, and SEQ ID NO: 141. The molecule can be any one of an antibody, an antibody fragment, a single chain variable fragment (scFv), a chimeric antigen receptor (CAR), a T cell receptor (TCR), a TCR mimic, a tandem scFv, a bispecific T cell engager, a diabody, a single-chain diabody (scDb), an scFv- Fc, a bispecific antibody, and a dual-affinity re-targeting antibody (DART). The molecule also can include an antigen-binding domain that can bind to an effector cell receptor selected from any one of CD3, CD28, CD4, CD8, CD16a, NKG2D, PD-1, CTLA-4, 4-1BB, 0X40, ICOS, CD27, and an Fc receptor. In some cases, the antigen-binding domain that can bind to an effector cell can bind to CD3, and the antigen-binding domain can include an amino acid sequence set forth in any one of SEQ ID NO : 170, SEQ ID NO : 171 , SEQ ID NO : 172, SEQ ID NO: 173, SEQ ID NO: 174, SEQ ID NO: 175, SEQ ID NO: 176, SEQ ID NO: 177, SEQ ID NO: 178, SEQ ID NO: 179, SEQ ID NO: 180, SEQ ID NO: 181, SEQ ID NO: 182, and SEQ ID NO:183.

In another aspect, this document features molecules that have an antigen-binding domain that can bind to a peptide-HLA-P2M complex, where the peptide can be derived from a modified RAS polypeptide. The modified RAS peptide can include from 7 amino acids to 25 amino acids ( e.g ., the modified RAS peptide can include 10 amino acids). The modified RAS peptide can include an amino acid sequence set forth in any one of SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4. In some cases, the modified RAS peptide can include the amino acid set forth in SEQ ID NO:2, and the antigen binding domain can include an amino acid sequence set forth in any one of SEQ ID NO: 142, SEQ ID NO: 143, SEQ ID NO: 144, SEQ ID NO: 145, SEQ ID NO: 146, SEQ ID NO: 147, SEQ ID NO: 148, and SEQ ID NO: 149. In some cases, the modified RAS peptide can include the amino acid sequence set forth in SEQ ID NO:3, and the antigen binding domain can include an amino acid sequence set forth in any one of SEQ ID NO: 150, SEQ ID NO: 151, SEQ ID NO: 152, SEQ ID NO: 153, SEQ ID NO: 154, SEQ ID NO: 155, SEQ ID NO: 156, SEQ ID NO: 157,

SEQ ID NO: 158, SEQ ID NO: 159, and SEQ ID NO: 160. In some cases, the modified RAS peptide can include the amino acid sequence set forth in SEQ ID NO:4, and the antigen binding domain can include an amino acid sequence set forth in any one of SEQ ID NO: 161, SEQ ID NO: 162, SEQ ID NO: 163, SEQ ID NO: 164, SEQ ID NO: 165, SEQ ID NO: 166,

SEQ ID NO: 167, SEQ ID NO: 168, and SEQ ID NO: 169. The molecule can be any one of an antibody, an antibody fragment, a scFv, a CAR, a TCR, a TCR mimic, a tandem scFv, a bispecific T cell engager, a diabody, a scDb, an scFv-Fc, a bispecific antibody, and a DART. The molecule also can include an antigen-binding domain that can bind to an effector cell receptor selected from any one of CD3, CD28, CD4, CD8, CD16a, NKG2D, PD-1, CTLA-4, 4- IBB, 0X40, ICOS, CD27, and an Fc receptor. In some cases, the antigen-binding domain that can bind to an effector cell can bind to CD3, and the antigen-binding domain can include an amino acid sequence set forth in any one of SEQ ID NO: 170, SEQ ID NO: 171, SEQ ID NO: 172, SEQ ID NO: 173, SEQ ID NO: 174, SEQ ID NO: 175, SEQ ID NO: 176, SEQ ID NO: 177, SEQ ID NO: 178, SEQ ID NO: 179, SEQ ID NO: 180, SEQ ID NO: 181, SEQ ID NO: 182, and SEQ ID NO: 183. In another aspect, this document features methods for treating a mammal having a cancer. The methods can include, or consist essentially of, administering to a mammal having cancer one or more molecules described herein ( e.g ., a molecule having an antigen binding domain that can bind to a peptide-HLA-P2M complex, where the peptide can be derived from a modified p53 polypeptide or a modified RAS polypeptide), where the cancer includes cancer cells expressing the modified peptide. The mammal can be a human. The cancer can be any one of Hodgkin’s lymphoma, non-Hodgkin’s lymphoma, acute myeloid leukemia, acute lymphoblastic leukemia, multiple myeloma, a myelodysplastic syndrome (MDS), a myeloproliferative disease, lung cancer, pancreatic cancer, gastric cancer, colorectal cancer, ovarian cancer, endometrial cancer, biliary tract cancer, liver cancer, breast cancer, prostate cancer, esophageal cancer, stomach cancer, kidney cancer, bone cancer, soft tissue cancer, head and neck cancer, glioblastoma multiforme, astrocytoma, thyroid cancer, germ cell tumor, and melanoma.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Although methods and materials similar or equivalent to those described herein can be used to practice the invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.

DESCRIPTION OF THE DRAWINGS

Figure 1 contains a graph showing flow cytometry on peptide-pulsed A2+ cells. T2 cells were peptide-pulsed overnight at 37°C in serum-free media with beta-2 microglobulin (b2M) protein only, or b2M with a p53 WTQ68-176) peptide (HMTEVVRRC; SEQ ID NO: 135), or p53 R175HQ68-176) peptide (HMTEVVRHC; SEQ ID NO:l). Cells were stained with 50 pL of phage supernatant per 50 pL of cells, washed and stained with rabbit anti -Ml 3 antibody, then washed and stained with anti-Rabbit-PE antibody. Cells were stained with a live/dead Near-infrared (IR) dye, washed, and analyzed by an iQue Screener (IntelliCyt, Albuquerque, NM).

Figure 2 contains a graph showing flow cytometry on peptide-pulsed A1+ cells. SigM5 cells were peptide-pulsed overnight at 37°C in serum-free media with b2M only, b2M with the H/K/N RAS WT(55-64) peptide (ILDTAGQEEY; SEQ ID NO: 136), b2M with a H/K/N RAS mutant Q61H(55-64) peptide (ILDTAGHEEY ; SEQ ID NO:2). Cells were stained with 50 pL of phage supernatant per 50 pL of cells, washed and stained with rabbit anti -Ml 3 antibody, then washed and stained with anti-Rabbit-PE antibody. Cells were stained with a live/dead Near-IR dye, washed, and analyzed by an iQue Screener (IntelliCyt, Albuquerque, NM).

Figure 3 contains a graph showing flow cytometry on peptide-pulsed A1+ cells. SigM5 cells were peptide-pulsed overnight at 37°C in serum-free media with b2M only, b2M with the H/K/N RAS WT(55-64) peptide (ILDTAGQEEY; SEQ ID NO: 136), b2M with a H/K/N RAS mutant Q61L(55-64) peptide (ILDTAGLEEY; SEQ ID NO:3). Cells were stained with 50 pL of phage supernatant per 50 pL of cells, washed and stained with rabbit anti -Ml 3 antibody, then washed and stained with anti-Rabbit-PE antibody. Cells were stained with a live/dead Near-IR dye, washed, and analyzed by an iQue Screener (IntelliCyt, Albuquerque, NM).

Figure 4 contains a graph showing flow cytometry on peptide-pulsed A1+ cells. SigM5 cells were peptide-pulsed overnight at 37°C in serum-free media with b2M only, b2M with the H/K/N RAS WT(55-64) peptide (ILDTAGQEEY; SEQ ID NO: 136), b2M with a H/K/N RAS mutant Q61R(55-64) peptide (ILDT AGREE Y ; SEQ ID NO:4). Cells were stained with 50 pL of phage supernatant per 50 pL of cells, washed and stained with rabbit anti -Ml 3 antibody, then washed and stained with anti-Rabbit-PE antibody. Cells were stained with a live/dead Near-IR dye, washed, and analyzed by an iQue Screener (IntelliCyt, Albuquerque, NM).

Figure 5 contains graphs showing that scDbs can induce mutation-specific T-cell cytokine release. p53 R175H(168-176)-A2 cl.2, cl.6, cl.15, cl.16, and cl.20 scDbs, containing the anti-CD3 clone UCHT1, were incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA-A2, p53(WT), p53(R175H), and GFP for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA. Figure 6 contains graphs showing that scDbs can induce mutation-specific T-cell cytokine release. H/K/N RAS Q61H(55-64)-Al cl.1, cl.2, and cl.4 scDbs, containing the anti- CD3 clone hUCHTlv9, were incubated at the specified concentrations with T cells and COS- 7 cells co-transfected with plasmids encoding various combinations of HLA-A1,

KRAS(WT), KRAS(Q61H), and GFP for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figure 7 contains graphs showing that scDbs can induce mutation-specific T-cell cytokine release. H/K/N RAS Q61L(55-64)-Al cl.1, cl.2, cl.9, and cl.13 scDbs, containing the anti-CD3 clone hUCHTlv9, were incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA-A1, KRAS(WT), KRAS(Q61L), and GFP for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figure 8 contains a graph showing scDb can induce mutation-specific T-cell cytokine release. H/K/N RAS Q61R(55-64)-Al cl.6 scDb, containing the anti-CD3 clone hUCHTlv9, was incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA-Al, KRAS(WT), KRAS(Q61R), and GFP for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figures 9A-9C contain graphs showing scDbs can react against Q61H, Q61L, and Q61R mutations in all 3 RAS isoforms, namely HRAS, KRAS, and NRAS. Fig. 9A H/K/N RAS mutant Q61H(55-64)-Al cl.l scDb was incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA- Al, H/K/N RAS(WT), H/K/N RAS(Q61H), H/K/N RAS(Q61K), H/K/N RAS(Q61L), and H/K/N RAS(Q61R) and for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA. Fig. 9B H/K/N RAS mutant Q61L(55- 64)-Al cl.2 scDb was incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA-Al, H/K/N RAS(WT), H/K/N RAS(Q61H), H/K/N RAS(Q61K), H/K/N RAS(Q61L), and H/K/N RAS(Q61R) and for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA. Fig. 9C H/K/N RAS mutant Q61R(55-64)-Al cl.6 scDb was incubated at the specified concentrations with T cells and COS-7 cells co-transfected with plasmids encoding various combinations of HLA-Al, H/K/N RAS(WT), H/K/N RAS(Q61H), H/K/N RAS(Q61K), H/K/N RAS(Q61L), and H/K/N RAS(Q61R) and for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figure 10 contains graphs showing scDb can induce mutation-specific T-cell cytokine release against tumor cell line. p53 R175H(168-176)-A2 cl.2 UCHT1 scDb was incubated at the specified concentrations with T cells and either with parental TYKnu or p53 knockout (KO) TYKnu for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figure 11 contains graphs showing scDb can induce mutation-specific T-cell cytokine release against tumor cell line. H/K/N RAS Q61L(55-64)-Al cl.2 UCHT1 scDb was incubated at the specified concentrations with T cells and either with parental HL-60 or HLA-A1 KO HL-60 for 20 hours at 37°C. Following co-culture, conditioned media was collected and assayed for secreted IFNy by ELISA.

Figure 12 shows that a MANAbody clone converted into a scDb can specifically kill tumor cells. p53 R175H(168-176)-A2 cl.2 UCHT1 scDb was incubated at the specified concentrations with T cells and either with parental TYKnu or p53 KO TYKnu for 20 hours at 37°C. Following co-culture, CellTiter-Glo was used to assay viable cells in each well. Percent target cell viability was calculated by subtracting the value from T cell only wells and normalizing to the value from target cell only wells.

Figure 13 shows that a MANAbody clone converted into a scDb can specifically kill tumor cells. H/K/N RAS Q61L(55-64)-Al cl.2 UCHT1 scDb was incubated at the specified concentrations with T cells and either with parental HL-60 or HLA-Al KO HL-60 for 20 hours at 37°C. Following co-culture, CellTiter-Glo was used to assay viable cells in each well. Percent target cell viability was calculated by subtracting the value from T cell only wells and normalizing to the value from target cell only wells.

Figures 14A-14C show biological and biophysical characteristics of scFv clone H2. (A) H2-scDb binding to immobilized p53^R175H/ HLA-A*02:01 (red) or p53^WT/HLA-A*02:01 (gray) pHLA was assessed by ELISA. Data shown represent mean ± SD of three technical replicates. (B) H2-scDb binding to p53^R175H/ HLA-A*02:01 was measured by single-cycle kinetics using SPR. H2-scDb was loaded at increasing concentrations, from 3, 12, 50, 200 to 800 nM. The blank- and reference-subtracted binding is shown for p53^R175H/HLA-A*02:01 (red) and p53^WT/HLA-A*02:01 (gray). H2-scDb binds to the p53^R175H/HLA-A*02:01 pHLA with a one-to-one binding kinetics at a KD of 86 nM (fitted line in black). There was negligible p53^WT/HLA-A*02:01 binding. (C) T2 cells pulsed with p53^R175H or p53^WT peptide was co-incubated with H2-scDb and T cells at an effector: target (E:T) ratio of 2: 1. IFN-g release was measured by ELISA (left) and cell lysis was evaluated by the CellTiter-Glo assay (right). Data indicate mean ± SD of three technical replicates and are representative of three independent experiments.

Figures 15A-15C show that H2-scDb activates T cells in the presence of tumor cells presenting p53^R175H. (A) HLA-A*02:01 positive tumor cell lines with different HLA expression levels and p53^R175H status were co-incubated with H2-scDb and T cells at an E:T ratio of 2:1. IFN-g release was measured by ELISA. Data indicate mean ± SD of 6 technical replicates and are representative of two independent experiments. The HLA-A*02 median fluorescence intensity (MFI) ratio is defined as MFI (anti-HLA-A*02)/MFI (isotype control). (B) Polyfunctional T-cell activation mediated by H2-scDb in response to KMS26 at an E:T ratio of 2: 1 was assessed by luminescence cytotoxicity and antibody -based assays (see Supplementary Materials). ECso (M) for each assay is shown in the corresponding graphs. Data indicate mean ± SD of three technical replicates and are representative of two independent experiments. (C) Real-time live-cell imaging of T cells with CellTracker Green CMFDA-labelled TYK-nu at an E:T ratio of 5: 1 with or without H2- scDb. Representative phase contrast and green fluorescence images taken at 24 hours (left) and 60 hours (right) after mixing cells are shown.

Figures 16A-16D show a determination of H2-scDb specificity using isogenic target cell lines. (A) HEK293FT and Saos-2 cell lines that were transfected with full-length p53^WT, full-length p53^R175H, or were not transfected were co-incubated with T cells at an E:T ratio of 2: 1 in the presence of increasing amounts of H2-scDb. IFN-g release was measured by ELISA. Data indicate mean ± SD of three technical replicates. (B) Cell lines expressing p53^R175H and transduced or not transduced with HLA-A*02:01 were co-incubated with T- cells and H2-scDb. IFN-g release was measured by ELISA. Experiments were performed at an E:T ratio of 2: 1 in duplicate (AU565) or triplicate (other 3 lines). (B) Cell lines expressing p53^R175H were transduced with HLA-A*02:01. Recognition of these cell lines by H2-scDb was assessed by IFN-g release. Experiments were performed at an E:T ratio of 2: 1 in duplicate (AU565) or triplicate (other 3 lines). Data indicate mean ± SD. (C) IFN-g release mediated by H2-scDb in response to parental tumor cell lines and their TP53 KO counterparts at an E:T ratio of 2: 1 (KMS26, TYK-nu) or 5 : 1 (KLE) was measured by ELISA. Data indicate mean ± SD of two (TYK-nu) or three (KMS26, KLE) technical replicates and are representative of two independent experiments. (D) Growth of parental (left) or TP53 KO (right) TYK-nu cells co-incubated with H2-scDb and T cells at an E:T ratio of 5:1 was measured by well confluence using real-time live-cell imaging. Data indicate mean ± SD of three technical replicates. One-way ANOVA with Tukey’s multiple comparison was used to evaluate statistical significance, *denotes E < 0.0001.

Figures 17A-17H show that H2 binds to the HLA-A*02:01 and the C-terminus of the p53^R175H neoantigen. (A) Overall structure of p53^R175H/HLA-A*02:01 bound to the H2-Fab fragment (PDB: 6W51). The HLA-A*02:01, b2 microglobulin (b2M), p53^R175H peptide, and the light and heavy chain of the H2-Fab are labeled accordingly. The p53^R175H nine amino acid peptide is between helices al and a2 of the HLA. (B) Structure of H2-Fab— p53^R175H/HLA-A*02:01 at 90° to that shown in (A). (C) Electron density map (2mFo-DFc) of the p53^R175H neoantigen. (D) Electron density map (2mFo-DFc) of a selected area of the H2-Fab at CDR-L3 from residues 95 to 99. (E) Zoom in of the interaction of H2-Fab to p53^R175H/HLA-A*02:01 with CDRs labeled in order from left to right: H2, HI, L3, H3, LI, L2. (F) Bird’s-eye view of surface representation of the HLA-A*02:01, p53^R175H peptide, and the contacting residues labeled according to CDRs of the H2-Fab. (G) Schematic representation of (F). (H) Diagram of the orientation angle of the H2-Fab. The angle of the orientation was calculated from two vectors: one from N and C termini of the peptide and the other between the intermolecular disulfide bonds of the VL and VH domains. The arrowed lines indicate the direction of the vectors.

Figures 18A-18F show a structural basis of H2 specificity and identification of putative cross-reactive peptides. (A) Detailed interactions of the p53^R175H neoantigen with HLA-A*02:01. The peptide (green) and the side chains (grey) of interacting residues of HLA-A*02:01 are represented as sticks. Hydrogen bonds are shown as dashed lines. (B) Perpendicular view of the p53^R175H peptide binding cleft. (C) C-terminus of the p53 peptide (aa Val6-Cys9) with Arg7 and His8 surrounded by the interacting residues of CDR-H1, -H2, -H3, and -L3 shown as sticks. Hydrogen bonds are shown as dashed lines. (D) T2 cells were loaded with 10 mM of HMTEVVRHC (SEQ ID NO: 1) peptide variants from the positional scanning library and co-incubated with 1 nM H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by cytometric bead array and the mean of triplicate wells was used to plot the heatmap. Black boxes represent the parental p53^R175H peptide. (E) Illustration of the binding pattern of H2-scDb as Seq2Logo graph (SEQ ID NO: 184), calculated by dividing the IFN-g value by 10⁴ and using the PSSM-Logo algorithm. (F) T2 cells were loaded with 10 mM of p53^R175H(SEQ ID NO:l), p53^WT(SEQ ID NO: 135), STAT2 (SEQ ID NO: 185), VP13A (SEQ ID NO: 186), or ZFP3 (SEQ ID NO:187) peptide and co incubated with 1 nM H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g secretion was measured by ELISA. Data indicate mean ± SD of three technical replicates.

Figures 19A-19C show in vivo antitumor efficacy of H2-scDb. In the early treatment model, NSG mice were engrafted with lxlO⁷ human T cells and either lxlO⁶ parental KMS26 (A) or lxlO⁶ TP53 KO KMS26 (B) on day 0. On day 1, intraperitoneal infusion pumps were placed to administer H2-scDb or isotype control scDb. (C) In the established tumor model, mice were engrafted with lxlO⁷ human T cells and 3.5xl0⁵ parental KMS26 on day 0, followed by administration of H2-scDb or isotype scDb at the specified dose on day 6. Tumor growth was monitored by bioluminescent imaging. N = 4 or 5 mice per group. Color bars denote the radiance (p/sec/cm²/sr) scale at each time point. Plotted data indicate mean ± SD. * denotes P <0.05 and NS denotes no statistical significance compared to isotype control by multiple t-test with Holm-Sidak correction.

Figure 20 shows the detection and quantification of p53^R175H neoantigen peptide in cells. COS-7 cells transfected with constructs expressing HLA-A*02:01 and p53^R175H, as well as cells lines with endogenous HLA-A*02:01 and p53^R175H expression, were analyzed for the presentation of p53^R175H neoantigen peptide HMTEVVRHC (SEQ ID NO:l; upper panel). Heavy isotope labeled p53^R175H neoantigen peptide was spiked into the assay and served as standards for absolute copy number quantification (lower panel).

Figure 21 shows a flow cytometric screening of phage clones enriched by panning. After 5 rounds of panning, phage clones from the enriched phage pool were isolated by limiting dilution and grown in deep 96-well plates. Supernatants containing individual phage clones were used to assess binding to T2 cells loaded with b2 macroglobulin (b2M) only, b2M plus p53^WT peptide, or b2M plus p53^R175H peptide via flow cytometry. The median fluorescence intensity (MFI) ratio was defined as MFI (R175H peptide)/MFI (WT peptide). NC, no phage control. Figure 22 contains a schematic representation of the structure of the T cell-engaging bispecific single-chain diabody (scDb) used in our experiments. VL, variable light domain; VH, variable heavy domain; pHLA, peptide-HLA complex; LL, long linker; SL, short linker.

Figure 23 shows a screening of scDb clones via IFN-g stimulation by p53-expressing cells. scDbs generated by linking each anti-p53^R175H/HLA-A*02:01 pHLA scFv clone with an anti-CD3 scFv (UCHT1) were coincubated with T cells and COS-7 cells transfected with GFP, HLA-A* 02:01 + GFP, HLA-A*02:01 + p53^WT, or HLA-A*02:01 + p53^R175H plasmids at an effectortarget (E:T) ratio of 1 : 1. After a 20-hour coincubation, the supernatant was harvested for IFN-g detection by ELISA. Arrows indicate clones H2 and H20. A2, HLA- A*02:01.

Figure 24 shows a characterization of H20-scDb. H20-scDb was incubated with biotinylated p53^R175H/HLA-A*02:01 (red) andp53^WT/HLA-A*02:01 (gray) pHLA monomers coated on streptavidin microplates at the specified concentrations, then binding detected with protein L and anti-protein L HRP. Data indicate mean ± SD of three technical replicates.

Figure 25 contains a comparison of scDbs generate with different anti-CD3 scFvs.

H2 was linked with different anti-CD3 scFvs in the scDb format and co-incubated with T cells and T2 cells pulsed with titrated concentrations of p53^R175Hor p53^WT peptide at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Data indicate mean ± SD of three technical replicates.

Figure 26 shows a H2-scDb-induced polyfunctional T-cell response. T-cell cytokine release mediated by H2-scDb in response to KMS26 cell line at an E:T ratio of 2: 1 was assessed by antibody-based assays. EC so (M) for each analyte is shown in the corresponding graphs. Data indicate mean ± SD of three technical replicates.

Figure 27 shows a flow cytometric evaluation of HLA-A*02 expression on tumor cell lines. Expression of HLA-A*02 on tumor cell lines was evaluated by flow cytometry. The red histogram represents staining with anti-HLA-A*02 (clone BB7.2), and the gray histogram represents staining with an isotype control. The median fluorescence intensity (MFI) ratio is defined as MFI (anti-HLA-A*02)/MFI (isotype control).

Figure 28 shows a flow cytometric evaluation of tumor cell lines transduced with HLA-A*02:01-encoding retrovirus. HLA-A*02:01 -encoding retrovirus was transduced into cell lines that weakly (AU565, SK-BR-3) or do not detectably (HuCCTl, CCRF-CEM) express HLA-A*02:01. Expression of HLA-A*02:01 in the parental and transduced and sorted cell lines was evaluated by flow cytometry. The red histogram represents staining with anti-HLA-A*02 (clone BB7.2) and the gray histogram represents staining with an isotype control.

Figures 29A-29B show a determination of H2-scDb specificity using CRISPR-edited isogenic cell lines. (A) Expression of p53 protein in the parental and TP53 KO clones of KMS26, KLE, and TYK-nu was assessed by Western blot staining with anti-p53 antibody (clone DO-1). (B) Cytotoxicity mediated by H2-scDb in response to parental tumor cell lines and their TP53 KO counterparts at an E:T ratio of 2: 1 (KMS26, TYK-nu) or 5: 1 (KLE) was measured by the Bio-Glo (KMS-26) or CellTiter-Glo (TYK-nu, KLE) assay. Data indicate mean ± SD of two (TYK-nu) or three (KMS26, KLE) technical replicates and are representative of two independent experiments.

Figures 30A-30B shows a H2-Fab-p53^R175H/HLA-A*02:01 complex. (A) Size- exclusion chromatogram of the pHLA-A*02:01 in complex with the H2-Fab. Protein was monitored by A280 nm with a major peak (-100 kDa). (B) Coomassie-stained gradient SDS-PAGE gel of the eluted fractions at 11-17 mL from (A).

Figures 31 A- 3 ID show that the neoantigen p53^R175H binds to HLA-A*02:01 in a canonical fashion. (A) Bird’s-eye view of the p53^R175H neoantigen interactions with HLA- A*02:01. The p53 peptide and the side chains of interacting residues of HLA-A*02:01 are represented as sticks. Hydrogen bonds are shown as dashed lines. The N-terminal Hisl is anchored by three tyrosine residues of HLA-A*02:01, one at the base of the cleft (Tyr7, not shown) and two on a2 (Tyrl59, 171), while its side chain is within hydrogen bonding distance of Lys66 (al) and Thrl63 (a2). Glu63 of HLA-A*02:01 al forms a hydrogen bond with the backbone amino of Metl69 at P2, an anchor residue of p53^R175H that is situated within the hydrophobic B pocket of the HLA. The main chain of Thr3 is stabilized by a hydrogen bond to Tyr99, located at the base of the cleft while the side chain of Glu4 forms a salt-bridge with the side chain of Arg65. Positions 5-8 (Vall72, Vall73, Argl74, Hisl75) of the p53^R175H neoantigen are stabilized by multiple hydrophobic and aliphatic residues with no direct hydrogen bonding contacts to the HLA-A*02:01. Towards the C-terminus of the neoantigen, the carboxyl group of Cysl76 at P9, another anchor residue that lies within the F pocket, is secured by Tyr84 (al) and Lysl46 (a2), while the side chain sulfhydryl is near Thrl43 on a.2. (B) Surface representation of the HLA-A*02:01 with the p53^R175H neoantigen shown as sticks. Anchor pockets B and F are circled in black. (C) Structural alignment of the following HLA-A*02:01-bound peptides in the binding pocket: p53^R175H (PDB ID 6W51), pi 049 (PDB ID 2JCC), NY-ESO-1 (PDB ID 3HAE), and WT1 (PDB ID 4WUU).

(D) Zoomed in view of (B) with helix al transparent and residues at position 7 (P7, Argl74) and 8 (P8, Hisl75) shown as sticks.

Figure 32 shows a hydrogen bonding pattern of the H2-Fab cage-like configuration. The imidazole ring of Hisl75 at P8 was at the center of the cage-like structure. The guanidinium group of Arg7 was at hydrogen bonding distance to the backbone carbonyl of Ala31 (CDR-H1). Another neoantigen-antibody direct contact involves the backbone carbonyl of Val6 hydrogen bonding with the side chain of Arg93 (CDR-L3).

Figures 33A-33P show a comparison of binding orientations among TCRm-pHLA and TCR-pHLA complexes. (A) Close depiction of binding of the H2-Fab to p53^R175H/HLA- A*02:01 with CDRs labeled as in Fig. 17. (B) Binding of a TCR to melanoma-associated antigen 3 (MAGE-A3) and HLA-A*01:01 (PDB 5BRZ). Same orientation as (A). The MAGE-A3 TCR displays the canonical, diagonal binding motif to that of most known TCR topologies. (C) Recognition of the 3M4E4 Fab for the NY-ESO-1157-165/HLA- A*02:01complex (PDB 3HAE). Same orientation as (A). (D) Binding of the ESK1 Fab to Wilms tumor 1 peptide and HLA-A*02:01 (PDB 4WUU). Same orientation as (A). (E, F, G and H) Bird’s-eye view of surface representation of the HLA-A*02:01/*01 :01 with the contacting residues of H2-Fab, MAGE- A3 TCR, 3M4E4 Fab, and ESK1 Fab, respectively, labeled and indicated with arrows. (I, J, K and L) Schematic representation of E, F, G, and H, respectively. H2-Fab-p53^R175H/HLA-A*02:01 shows a different mode of antibody recognition compared with other Fab/TCR-pHLA complexes. (M, N, O and P) Schematic representation of Fab/TCR orientation angle.

Figure 34 shows binding of p53^R175H positional scanning library peptides to HLA- A*02:01. A peptide library was generated by systemically substituting the amino acid at each position of the target peptide (HMTEVVRHC; SEQ ID NO: 1) with each of the remaining 19 common amino acids. T2 cells were loaded with each of the variant peptides at lOOuM in the presence of 10 pg/ml b2M and anti-HLA-A*02 antibody (clone BB7.2). Peptide binding was evaluated by flow cytometry. Black boxes represent the parental peptide. MFI, median fluorescence intensity.

Figure 35 shows recognition of the p53^R175H positional scanning library peptides by H2-scDb. T2 cells loaded with variant peptides in the positional scanning library was incubated with InM H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Dotted lines represent 20% of parental peptide IFN-g value. Peptides for Positions 1-9 are SEQ ID NOs: 188-196, respectively. The binding motif established by the 20% reactivity cutoff, expressed in PROSITE pattern, was x-[AåLMVNQTC]-[ST]-[DE]- [IV] - [IM V S T] -R-H- [ AIL V GHS T Y C ] (SEQ ID NO: 197). Data indicate mean ± SD of three technical replicates.

Figures 36A-36B contain an assessment of H2-scDb cross-reactivity H2-scDb was co-incubated with T cells and COS-7 cells transfected with HLA-A*02:01 and full-length p53^R175H, STAT2, orZFP3 at an E:T ratio of 5:1. (A) Expression of target proteins by COS- 7 cells was assessed by Western blot staining. (B) IFN-g secretion was measured by ELISA. The signals for all of the transfectants except for p53^R175H were indistinguishable and are clustered near the x-axis. Data indicate mean ± SD of three technical replicates.

Figure 37 shows the body weight of NSG mice in the established tumor model. NSG mice were engrafted with lxlO⁷ human T cells and 3.5xl0⁵ parental KMS26 on day 0, followed by administration of the specified scDb on day 6. Body weight of the mice was serially monitored. N = 5 mice per group. Data shown represent mean ± SD.

Figures 38A-38H. ELISA and flow cytometry characterization of RAS MANA scFvs. scFvs against RAS MANAs were characterized using ELISA, flow cytometry, and SPR. (A,D-F) Biotinylated G12V or G12WT pHLA-A3 (A) or Q61WT, Q61H, Q61L, or Q61R pHLA-Al (D-F) were coated on a streptavidin plate at the specified concentrations. Recombinant RAS G12V clone V2 (A), Q61H clone HI (D), Q61L clone L2 (E), or Q61R clone R6 (F) scFvs were incubated in the wells at 1 pg/mL, followed by detection with protein L and horseradish peroxidase (HRP)-conjugated anti-protein L. All ELISAs were performed in triplicate. (B,G) T2A3 or SigM5 cells were pulsed with the specified peptides at 50 mM, followed by flow cytometric analysis. (B) T2A3 cells were stained with V2 scFv pre-conjugated to anti-FLAG-phycoerythrin (PE), with mean fluorescence intensities (MFI) plotted. (G) SigM5 cells were stained with clone HI, L2, or R6 phage, then detected with rabbit anti -Ml 3 phage and PE-conjugated anti-rabbit antibodies. Phage background- subtracted PE MFI are plotted. (C,H) V2 scFv and L2 single chain diabody (scDb) binding were evaluated by single-cycle kinetics using SPR. (C) The V2 scFv bound to G12V pHLA- A3 with one-to-one binding kinetics and a KD of 8.7 nM, with negligible binding to G12WT pHLA-A3. (H) The L2-U scDb bound to the Q61L pHLA-Al with one-to-one binding kinetics and a KD of 65 nM, with negligible binding to Q61WT pHLA-Al.

Figures 39A-39E. Schematic and ELISA characterization of MANA scDbs. (A) Schematic showing the optimal bispecific format, an scDb with variable light (VL) and variable heavy (VH) domains arranged in the following order: VLV2-VHUCHTI-VLUCHTI- VHV2. SL, short linker (GGGGS; SEQ ID NO:200); LL, long linker (GGGGS)₃ (SEQ ID NO:201). (B-E) anti-MANA/anti-CD3 scDbs were characterized via ELISA. Biotinylated pHLA-A3, pHLA-Al or recombinant CD3e/5 protein were coated on a streptavidin plate. Recombinant V2-U (B), Hl-U (C), L2-U (D), or R6-U (E) scDbs were incubated at the specified concentrations then detected with protein L and anti-protein L HRP. All ELISAs were performed in triplicate.

Figures 40A-40D. Co-cultures with peptide-pulsed cells. (A,B) T2A3 cells were pulsed with either the G12V or G12WT peptides at the specified concentrations. (C,D) SigM5 cells were pulsed with either the Q61L or Q61WT peptides at the specified concentrations. 5 x 10⁴ (T2A3) or 2.5 x 10⁴ (SigM5) peptide-pulsed cells were combined with 5 x 10⁴ human T cells (effectontarget ratio or E:T = 1:1 or 2: 1) and V2-U (A,B), V2-U2 scDb (A,B), or L2-U scDb (C,D) at 1 nM. Plates were incubated for 24 hours and the supernatants assayed for IFNy (A,C). Target cell cytotoxicity was assayed using CellTiter- Glo (B,D). Percent cytotoxicity was determined by subtracting the T cell signal and normalizing to the no peptide condition. All experiments were performed in triplicate.

Figures 41A-41D. Co-culture with transfected COS-7 cells. COS-7 cells were transfected with a 1 : 1 ratio of plasmids encoding HLA-A3 (“A3”) or HLA-A1 (“Al”) and RAS variants or other negative controls. 24 hours later, 1 x 10⁴ COS-7 cells were combined with 5 x 10⁴ human T cells (E:T = 5:1) and V2-U (A), Hl-U (B), L2-U (C), or R6-U (D) scDb at the specific concentrations. Plates were incubated for 24 hours and supernatants assayed for IFNy. All experiments were performed in triplicate.

Figures 42A-42F. Effects of V2-U scDb on co-cultures of human T cells with cancer cells. 2 x 10⁴ target cells from parental NCI-H441 (A,B), NCI-H441 isogenic variants (C,D) or NCI-H358 isogenic variants (E,F) were combined with 6 x 10⁴ human T cells (E:T = 3:1) and V2-U scDbs at the specified concentrations. Cells were incubated for 24 hours and assayed for IFNy release (A,C,E) and target cell cytotoxicity using CellTiter-Glo (B,D,F). Cytotoxicity was calculated by subtracting the T cell alone signal and normalizing to the no scDb condition (considered as 0% cytotoxicity). All experiments were performed in triplicate. For KRAS genotypes: V/D, G12V/frameshift (also see Figure 55).

Figures 43A-43E. Effects of L2-U scDb on co-cultures of human T cells with cancer cells. 2.5 x 10⁴ target cells from different cancer cell lines expressing HLA-A1, RAS Q61L, or both (A), parental HL-60 cells (B-C), or HL-60 isogenic variants expressing different RAS Q61 mutations or with HLA-A1 knocked out (KO) (D-E) were combined with 5 x 10⁴ human T cells (E:T = 2: 1) and L2-U scDb at the specified concentrations. Cells were incubated for 24 hours and assayed for IFNy release by ELISA (A,B,D) and target cell cytotoxicity (C,E). Target cell cytotoxicity was assessed via CellTiter-Glo. Cytotoxicity was calculated by subtracting the T cell alone signal and normalizing to the no scDb condition (considered as 0% cytotoxicity). All experiments were performed in triplicate. (For HL-60 NRAS genotypes, also see Figure 55).

Figures 44A-44D. Peptide scanning to assess V2-U and L2-U scDb cross-reactivity. Each amino acid position of the G12V and Q61L 10-mer peptides were systematically changed to the other 19 amino acids, thereby generating libraries of variant peptides each differing from the original peptide by a single amino acid. T2A3 cells were pulsed with 10 mM of the G12V peptide scanning library (A,C) and SigM5 cell cells were pulsed with 10 mM of the Q61L peptide scanning library (B,D). 2.5 x 10⁴ peptide-pulsed target cells were combined with 5 x 10⁴ human T cells (E:T = 2: 1) and either the V2-U (A,C) or L2-U scDb (B,D) at 1 nM. (A,B) Plates were incubated for 24 hour and assayed for IFNy release, with the mean of three technical replicates plotted as a heat map. Black boxes indicate amino acids in the parental peptides. (C,D) Illustration of the binding pattern of V2-U and L2-U scDb as Seq2Logo graph (SEQ ID NOs:683 and 684, respectively), calculated by dividing the IFNy value by 10³ and using the PSSM-Logo algorithm.

Figures 45A-45C. KRAS neoantigen transitions detected through MANA-SRM. (A- C) COS-7 cells transfected with constructs expressing HLA-A*03:01 and KRAS G12V (A), as well as cells lines NCI-H441 (B) and CFPAC-1 (C) with endogenous HLA-A*03:01 and KRAS G12V expression, were analyzed for the presentation of G12V[7-16] VVVGAVGVGK (SEQ ID NO:205; left) and G12V[8-16] VVGAVGVGK (SEQ ID NO:206; right) peptides (lower panels). The presence or absence of these peptides is denoted with a red arrow or “X,” respectively. Heavy isotope labeled RAS G12V peptides were spiked into the assay and served as standards for absolute copy number quantification (upper panels). Transition details for each peptide in each sample were shown on the right side of each individual plot after zooming in on the retention time (x-axis). Peptide quantification was performed based on the plots as previously described.

Figures 46A-46C. Design and sequencing of CDR-H3 of the phage library. (A) Expected amino acid diversity at variable codons of CDR-H3, using Rabat numbering. (B) Expected percent of clones with a given codon present in CDR-H3. (C) Expected vs. actual amino acid diversity at variable codons in CDR H3. The actual amino acid diversity was determined by MiSeq next-generation sequencing (NGS) of a portion of the library and subsequent analysis of the top 100,000 most frequent reads. T, theoretical distribution; M, MiSeq NGS analysis.

Figures 47A-47E. Characterization of the V2 scFv. (A) ELISA with monoclonal phage. After 4, 5, or 6 rounds of selection, monoclonal phage were amplified in bacteria in a 96-well plate format. Well H12 was not inoculated and thus served as a no phage control. Monoclonal phage were incubated in G12V pHLA-A3 or G12WT pHLA-A3 coated streptavidin ELISA plate. Plates were washed and phage were detected using rabbit anti- M13 and HRP-conjugated anti-rabbit antibodies. Phage clone V2 was identified in the four wells designated by the red arrows. (B) Flow cytometry. T2A3 cells pulsed or not pulsed with the indicated peptides were incubated with the top four candidate phage clones selected for their ability to bind G12V pHLA-A3. The binding was assessed by rabbit anti-phage Ml 3 and PE-conjugated anti -rabbit antibodies. MFI are plotted. (C) ELISA with V2 phage after dilution. Biotinylated pHLA-A3 were coated on a streptavidin plate. V2 phage were incubated at the specified dilutions and detected as in (A). (D) ELISAs with V2 scFv. Similar to (C), except that recombinant V2 scFv was used instead of phage, and detection employed protein L and HRP-conjugated anti-protein L. All ELISAs were performed in triplicate. (E) Peptide pulsing of T2A3 cells. T2A3 cells were pulsed with the specified peptides at 50 mM. Cells were stained with anti-HLA-A3 monoclonal antibody GAP. A3 conjugated to PE. MFI are plotted.

Figures 48A-48F. Characterization of the RAS Q61H, Q61L, and Q61R scFvs. (A) Flow cytometry. After 5 rounds of selection, monoclonal phage were amplified in bacteria in 96-well plate format. Monoclonal phage were sequenced and clustered to unique phage clones. (A,C,E) SigM5 cells pulsed with the specified peptides were incubated with phage clones from Q61H, Q61L, and Q61R selections, respectively and assessed by rabbit anti- phage Ml 3 and PE-conjugated anti -rabbit antibodies. Mean fluorescence intensities (MFI) are plotted. The top candidates are indicated with arrows. (B,D,F) COS-7 cells were transfected with a 1:1 ratio of plasmids encoding HLA-A1 (“Al”) and KRAS Q61 variants (or other negative controls). 24 hours later, 5 x 10⁴ COS-7 cells were combined with 5 x 10⁴ human T cells (effectortarget ratio or E:T = 1:1) and single chain diabody (scDb) proteins generated from the top flow cytometry candidates at the specified concentrations. Plates were incubated for 24 hours and assayed for secreted IFNy. Black arrows denote the single most specific and reactive clone for each target pHLA complex selected for further investigation.

Figures 49A-49D. Bispecific antibody formats and format comparison. (A) Cartoon depictions of the six bispecific formats tested with the RAS G12V pHLA-A3 targeting V2 scFv. (B) Schematic showing the tested orientations of the variable light (VL) and variable heavy (VH) domains of the V2 scFv with different anti-CD3 scFv clones. SL, short linker (GGGGS; SEQ ID NO:200); ML, medium linker (GGGGS)₂ (SEQ ID NO:694); LL, long linker (GGGGS)3 (SEQ ID NO:201). (C) ELISAs of V2-bispecific antibody formats to assess binding to G12V pHLA-A3 and recombinant CD3e/5 protein. Biotinylated G12V pHLA-A3 or CD3e/5 protein was coated on a streptavidin plate. Different bispecific antibody formats were incubated in wells at specified concentrations followed by detection with protein L and HRP-conjugated anti-protein L. (D) Testing of V2 bispecific formats in co-cultures of T cells and peptide-pulsed T2A3s. T2A3 cells were pulsed with either the G12V or G12WT peptide at the specified concentrations. 2.5 x 10⁴ peptide-pulsed T2A3 cells were combined with 5 x 10⁴ T cells (E:T = 2: 1) and V2 bispecific formats at either 1 nM or 0.2 nM bispecific antibody concentration. Plates were incubated for 24 hours and assayed for IFNy release. Note that the scFv-Fc format is a heterodimer of an FcKnob and an FcHole protein, with the bivalent scFv-Fc containing one V2 moiety and one anti-CD3 moiety, and the trivalent scFv-Fc containing two V2 moieties and one anti-CD3 moiety.

Figures 50A-50D. V2 scDbs made with various anti-CD3 clones. Twelve different anti-CD3 clones were tested in the VLV2-VHCD3-VLCD3-VHV2 format. (A) scDbs (all bearing C -terminal 6xHIS tags) were expressed in 293FT cells and purified identically. See Table 12 for sequences of the anti-CD3 clones. The Western blot shows the purified scDbs detected with anti-6xHIS and HRP-conjugated anti-rabbit antibodies. (B) Biotinylated recombinant

CD3e/6 protein was coated on a streptavidin plate. V2 scDbs were added to the plate at 2 pg/ml and detected with protein L and HRP-conjugated anti-protein L. (C) COS-7 cells were transfected with a 1 : 1 ratio of plasmids encoding HLA-A3 and KRAS G12V. 24 hours later, 1 x 10⁴ COS-7 cells were combined with 5 x 10⁴ human T cells (E:T = 5:1) and V2 scDbs at the indicated concentrations. Cells were incubated for 24 hours and assayed for secreted IFNy. (D) 2 x 10⁴ NCI-H441 target cells were combined with 6 x 10⁴ human T cells (E:T = 3:1) and V2 scDb proteins at the indicated concentrations. Cells were incubated for 24 hours and assayed for secreted IFNy

Figures 51 A- 5 IE. Specificity of scDbs. (A) Biotinylated pHLA-A3 or biotinylated recombinant CD3e/6 protein were coated on a streptavidin plate. V2-U2 scDb was added to the plate at the indicated concentrations and detected with protein L and HRP-conjugated anti-protein L. Each ELISA was performed in triplicate. (B) Biotinylated pHLA-A3 were coated on a streptavidin plate. ScDb was incubated with pHLA at the indicated concentrations, then incubated with recombinant CD3e/6 containing a human Fc tag and detected with an HRP-conjugated anti-human Fc antibody. (C-E) Biotinylated RAS Q61 pHLA-Al (Q61WT, Q61H, Q61L, or Q61R) were coated on streptavidin plates at the indicated concentrations. Recombinant Hl-U (C), L2-U (D), or R6-U (E) scDb was added to the wells at 10 nM and detected with protein L and HRP-conjugated anti -protein L.

Figures 52A-52N. Additional peptide-pulsing data. (A) T2A3 cells were pulsed with the RAS G12V or G12WT peptides at the indicated concentrations. 5 x 10⁴ T2A3 peptide- pulsed cells were co-cultured with 5 x 10⁴ human T cells (E:T = 1 : 1) in the presence of either

V2-U or V2-U2 scDb at 1 nM. Cells were incubated for 24 hours and TNFa secretion measured. (B) Quantification of G12V peptide on the surface of T2A3 cells. T2A3 cells were pulsed with the G12V peptide at the specified concentrations. G12V pHLA complexes on the cell surface were quantified using the V2 scFv and Quantibrite- and Quifikit-based methods. (C,D) G12 peptide-pulsed dendritic cells. HLA-A3+ immature dendritic cells

(iDCs) were pulsed with either the G12V or G12WT peptide at the specified concentrations.

1 x 10⁴ peptide-pulsed iDCs cells were combined with 5 x 10⁴ human T cells (E:T = 5:1) and either V2-U or V2-U2 scDb at 1 nM. Plates were incubated for 24 hours and assayed for secreted IFNy (C) and TNFa (D). (E-H) Q61 peptide-pulsed SigM5 cells. SigM5 cells were pulsed with the RAS Q61H, Q61R, or Q61WT peptide at the indicated concentrations. 2.5 x

10⁴ peptide-pulsed SigM5 cells were combined with 5 x 10⁴ human T cells (E:T = 2:1) and either Hl-U (E,F) or R6-U (G,H) scDb at 1 nM. Cells were incubated for 24 hours and assayed for secreted IFNy (E,G). Target cell cytotoxicity was assayed using CellTiter-Glo (F,H). (I-N) Q61 peptide-pulsed dendritic cells. HLA-A1+ iDCs were pulsed with either the Q61H, Q61L, Q61R, or Q61WT peptide at the indicated concentrations. 1 x 10⁴ peptide- pulsed iDCs cells were combined with 5 x 10⁴ human T cells (E:T = 5:1) and Hl-U (I,L), L2- U (J,M), or R6-U (K,N) scDb at 1 nM. Cells were incubated for 24 hours and assayed for secreted IFNy (I-K) or TNFa (L-N). All experiments were performed in triplicate.

Figures 53 A-53D. Western blots and co-cultures with transfected COS-7 cells. COS- 7 cells were transfected with a 1 : 1 ratio of plasmids encoding HLA-A3 (“A3”) and KRAS variants (or other negative controls). 24 hours later, cells were harvested for western blots (A) or co-cultures (B-D). (A) Western blots for KRAS and HLA-A3 in transfected COS-7. Cells were pelleted and snap frozen and analyzed via western for KRAS, HLA-A3, and b- actin protein expression. (B-D) 1 x 10⁴ COS-7 cells were combined with 5 x 10⁴ human T cells (E:T = 5:1) and V2-U (B) or V2-U2 scDb (C,D) at the indicated concentrations. Cells were incubated for 24 hours and supernatant assayed for secreted IFNy (C) or TNFa (B,D). All ELISAs were performed in triplicate.

Figure 54. HLA-A3 expression in target cell lines. 5 x 10⁵ target cells were stained with PE-conjugated anti-HLA-A3 clone GAP. A3 (shaded red) or PE-conjugated mouse isotype IgG2a control (shaded gray). Histograms show PE intensities of viable cells.

Figure 55. Sanger sequencing of RAS alleles in CRISPR-modified cell lines. Sanger sequencing of genomic DNA from the KRAS locus of NCI-H441 parental and G13D-KI clones (SEQ ID NO:695), from the KRAS locus of NCI-H358 parental and G12V-KI clones (SEQ ID NO:696), and from the NRAS locus of HL-60 parental, Q61H-KI, and Q61R-KI clones (SEQ ID NO:697). Note that KRAS codons are shown in the antisense orientation.

Figures 56A-56B. Effects of V2-U2 scDb on co-culture of T cells withNCI-H441 isogenic cell lines. 2 x 10⁴ parental NCI-H441 cells or its isogenic clones were combined with 6 x 10⁴ human T cells (E:T = 3:1) and V2-U2 scDb at the indicated concentrations.

Cells were incubated for 24 hours and assayed for IFNy release (A) or target cell cytotoxicity with CellTiter-Glo (B). All experiments were performed in triplicate.

Figure 57. Poly-functional immune response elicited by V2-U scDb. V2-U scDb at the indicated concentrations was incubated with 2.5 x 10⁴ NCI-441 cells and 5 x 10⁴ human T cells (E:T = 2: 1). Cells were incubated for 24 hours and assayed for secreted IFNy, TNFa, IL-2, granzyme B, and perforin using Luminex beads. All experiments were performed in triplicate.

Figures 58A-58B. Effects of V2-U2 scDb on co-culture of T cells withNCI-H358 isogenic cell lines. 2 x 10⁴ parental NCI-H358 cells or its isogenic clones were combined with 6 x 10⁴ human T cells (E:T = 3:1) and V2-U2 scDb at the indicated concentrations.

Cells were incubated for 24 hours and assayed for IFNy release (A) or target cell cytotoxicity with CellTiter-Glo (B). All experiments were performed in triplicate.

Figures 59A-59B. Effects of V2 scDbs on IFNy secretion from T cells in co-cultures with HLA-A3+ cell lines. 2 x 10⁴ target cells were combined with 6 x 10⁴ human T cells (E:T = 3 : 1) and V2-U (A) or V2-U2 (B) scDb at the indicated concentrations. The cells were incubated for 24 hours and assayed for IFNy release. All experiments were performed in triplicate. *** denotes P < 0.001 and NS denotes no statistical significance between CFPAC- 1 and each of the other cell lines without the RAS G12V mutation, as analyzed by two-way ANOVA collapsed across different scDb concentrations for a given cell line, with Tukey’s correction for multiple comparisons.

Figure 60. HLA-A1 expression in target cell lines. 5 x 10⁵ target cells were incubated with anti-HLA-Al/Al 1/A26 clone 8.L.101 (shaded red) or mouse IgM isotype control (shaded gray), then stained with a PE-conjugated anti-mouse antibody. Histograms show PE intensities of viable cells. Figure 61. Poly-functional immune response elicited by L2-U scDb. L2-U scDb at the indicated concentrations was incubated with 2.5 x 10⁴ HL-60 cells and 5 x 10⁴ human T cells (E:T = 2: 1). Cells were incubated for 24 hour and assayed for secreted IFNy, TNFa, IL- 2, granzyme B, and perforin using Luminex beads. All experiments were performed in triplicate. Figures 62A-62F. Testing potential cross-reactive peptides. (A) T2A3 cells were pulsed with 50 mM of the indicated peptides. Pulsed cells were incubated with V2 phage, followed by staining with rabbit anti -Ml 3 phage and PE-conjugated anti-rabbit antibodies or incubated with the PE-conjugated anti-HLA-A3 monoclonal antibody GAP. A3. PE MFI are plotted. (B) Western blot showing endogenous Rab-7b expression in PBMCs, monocytes (Mono), immature DCs (iDC), mature DCs (mDC), and Hs 695T cells, and overexpression of Rab-7b in HLA-A3 and RAB7B co-transfected COS-7 and HCT 116 versus HLA-A3 and GFP co-transfected controls. (C) 1 x 10⁴ NCI-H441 cells or 5 x 10⁴ PBMCs, Mono, iDC, or mDC, with or without pulsing with the G12V peptide, were combined with 5 x 10⁴ human T cells (E:T = 5: 1 or 1 : 1) and the V2-U scDb at 1 nM. Cells were incubated for 24 hours and assayed for secreted IFNy. Normal human cells were derived from an HLA-A3+ donor. (D) Hs 695T were transfected with HLA-A3 or HLA-A2 (negative control) encoding plasmids to assess for endogenous presentation and V2-U scDb recognition of the Rab-7b peptide. As a positive control, HI -A J-transfected cells were pulsed with the G12V peptide or dimethylformamide (DMF, solvent) only. Parental NCI-H441 cells were included as a positive control and NCI-H441 HLA-A3- KO (A3-KO) and NCI-H441 (KRAS G13D/WT) clone 1 (G13D-KI) were included as negative controls. In each well, 2 x 10⁴ target cells were combined with 5 x 10⁴ human T cells (E:T = 5:2) and V2-U scDb at the indicated concentrations. Plates were incubated for 24 hours and assayed for secreted IFNy. (E,F) COS-7 (E) and HCT 116 (F) cells were transfected with either vector only (GFP) or a 1:1 ratio of plasmids encoding HLA-A3 and vector only (GFP), full length KRAS WT, KRAS G12V, or Rab-7b. In each well, 2 x 10⁴ target cells were combined with 5 x 10⁴ human T cells (E:T = 5:2) and V2-U scDb at the indicated concentrations. Cells were incubated for 24 hour and assayed for secreted IFNy.

Figures 63A-63C. Positional scanning of target peptides that could potentially react with V2 or L2 scDb. (A) Each amino acid of the G12V peptide was systematically changed to the other 19 amino acids. T2A3 cells were pulsed with 10 mM of the G12V peptide library. 2.5 x 10⁴ peptide-pulsed T2A3 cells were combined with 5 x 10⁴ human T cells (E:T = 2: 1) and V2-U scDb at 1 nM. Cells were incubated for 24 hour and assayed for secreted IFNy. (B) Each amino acid of the Q61L peptide was systematically changed to the other 19 amino acids. SigM5 cells were pulsed with 10 pM of the Q61L peptide library. 2.5 x 10⁴ peptide-pulsed SigM5 cells were combined with 5 x 10⁴ human T cells (E:T = 2:1) and L2-U scDb at 1 nM. Cells were incubated for 24 hours and assayed for secreted IFNy. (C) SigM5 cells were pulsed with the Q61L, Q61WT, or CHD4 peptide. 2.5 x 10⁴ peptide-pulsed SigM5 cells were combined with 5 x 10⁴ human T cells (E:T = 2: 1) and L2-U scDb at 1 nM. Cells were incubated for 24 hours and assayed for secreted IFNy.

Figures 64A-64B. L2-U effects on tumor growth in mouse model systems. (A, B) Mice were engrafted with 1 x 10⁷ human T cells and 5 x 10⁵ luciferase-expressing parental HL-60 (A) or CRISPR-edited HL-60 (B) on day 0. One day later, after tumor engraftment was established by bioluminescent imaging, mice were randomized with respect to tumor burden and intravenously injected with 1 x 10⁷ human T cells. They were then immediately implanted with osmotic pumps delivering L2-U or isotype scDb (V2-U scDb) at 0.42g/kg/day. Tumor growth was monitored by bioluminescent imaging. N = 7 mice per group. Plotted data represent mean ± SD. * denotes P <0.05 and NS denotes no statistical significance compared to isotype control according to multiple t-tests with Bonferroni-Dunn correction.

Figures 65A-65B. Body weights of NSG mice treated with scDbs. (A,B) NSG mice were implanted with tumors and treated as described in Figure 64. Body weights of the mice were serially monitored. Data shown represent mean ± SD.

Figure 66. Diagram of scFv phage library phagemid. Oligonucleotides encoding the scFv and synthesized using TRIM technology were incorporated into a pADL-lOb phagemid. This phagemid contains an FI origin, a transcriptional repressor to limit un-induced expression, a lac operator, and a lac repressor. The scFv was synthesized with a pelB periplasmic secretion signal and was subcloned downstream of the lac operator. A linker (GGGSGGGGSGGGAS; SEQ ID NO:698) connects the variable light and heavy chains of the scFv. A FLAG (DYKDDDDK; SEQ ID NO: 190) epitope tag was placed immediately downstream of the variable heavy chain, which was followed in frame by the full-length Ml 3 pill coat protein sequence.

Figures 67A-67C. Biological and biophysical characteristics of scFv clone H2. (A) H2-scDb binding to immobilized p53^R175H/ HLA-A*02:01 (red) or p53^WT/HLA-A*02:01 (gray) pHLA was assessed by ELISA. Data shown represent mean ± SD of three technical replicates. (B) H2-scDb binding to p53^R175H/ HLA-A*02:01 was measured by single-cycle kinetics using SPR. H2-scDb was loaded at increasing concentrations, from 3, 12, 50, 200 to 800 nM. The blank- and reference-subtracted binding is shown for p53^R175H/HLA-A*02:01 (red) and p53^WT/HLA-A*02:01 (gray). H2-scDb binds to the p53^R175H/HLA-A*02:01 pHLA with a one-to-one binding kinetics at a KD of 86 nM (fitted line in black). There was negligible p53^WT/HLA-A*02:01 binding. (C) T2 cells pulsed with p53^R175Hor p53^WT peptide were co-incubated with H2-scDb and T cells at an effectontarget (E:T) ratio of 2: 1. IFN-g release was measured by ELISA (left) and cell lysis was evaluated by the CellTiter-Glo assay (right). Data indicate mean ± SD of three technical replicates and are representative of three independent experiments. Figures 68A-68D. H2-scDb activates T cells in the presence of tumor cells presenting p53^R175H. (A) Illustration depicting the mechanism of action of H2-scDb. (B) HLA-A*02:01 positive tumor cell lines with different HLA expression levels and p53^R175H status were co incubated with H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Data indicate mean ± SD of six technical replicates and are representative of two independent experiments. The HLA-A*02 median fluorescence intensity (MFI) ratio is defined as MFI (anti-HLA-A*02)/MFI (isotype control). (C) Poly functional T-cell activation mediated by H2-scDb in response to KMS26 at an E:T ratio of 2: 1 was assessed by luminescent cytotoxicity and antibody-based assays (see Supplementary Materials). ECso (M) for each assay is shown in the corresponding graphs. Data indicate mean ± SD of three technical replicates and are representative of two independent experiments. (D) Real-time live-cell imaging of T cells with GFP-labelled TYK-nu at an E:T ratio of 5:1 with or without H2-scDb. Representative phase contrast and green fluorescence images taken at 24 hours (top) and 96 hours (bottom) after mixing cells are shown.

Figures 69A-69E. Determination of H2-scDb specificity using isogenic target cell lines. (A) Methods of the generating of isogenic cell line pairs in cells with different HLA and p53 backgrounds. (B) HEK293FT and Saos-2 cell lines that were transfected with full- length p53^WT, full-length p53^R175H, or were not transfected were co-incubated with T cells at an E:T ratio of 2:1 in the presence of increasing amounts of H2-scDb. IFN-g release was measured by ELISA. Data indicate mean ± SD of two technical replicates. (C) Cell lines expressing p53^R175H and transduced or not transduced with HLA-A*02:01 were co-incubated with T-cells and H2-scDb. IFN-g release was measured by ELISA. Experiments were performed at an E:T ratio of 2: 1 in three technical replicates. (D) IFN-g release mediated by H2-scDb in response to parental tumor cell lines and their TP 53 KO counterparts at an E:T ratio of 2: 1 (KMS26, TYK-nu) or 5 : 1 (KLE) was measured by ELISA. Data indicate mean ± SD of two (TYK-nu) or three (KMS26, KLE) technical replicates and are representative of two independent experiments. * /^J<0.05, **P <0.01, ***p <0.001 by two-tailed t-test. (E) Parental (left) or TP53 KO (right) TYK-nu cells labeled with nuclear GFP were co-incubated with H2-scDb and T cells at an E:T ratio of 2: 1 was measured by real-time live-cell imaging. Data indicate mean ± SEM of twelve technical replicates. One-way ANOVA with Tukey’s multiple comparison was used to evaluate statistical significance, ****denotes P < 0.0001. Figures 70A-70H. H2-Fab binds to the HLA-A*02:01 and the C-terminus of the p53^R175H neoantigen. (A) Overall structure of p53^R175H/HLA-A*02:01 bound to the H2-Fab fragment (PDB ID 6W51). HLA-A*02:01 and b2 microglobulin (b2M) are colored in gray and gold, respectively. The H2-Fab is colored according to the heavy (blue) and light (cyan) chains of the Fab fragment. The p53^R175H nine amino acid peptide is shown in light green between helices al and a2 of the HLA. (B) Structure of H2-Fab-p53^R175H/HLA-A*02:01 at 90° to that shown in (A). (C) Electron density map (2mFo-DFc) of the p53^R175H neoantigen contour at 1s. (D) Electron density map (2mFo-DFc) of a selected area of the H2-Fab at CDR-L3 from residues 95 to 99 contoured at 1s. (E) Zoom in of the interaction of H2-Fab to p53^R175H/HLA-A*02:01 with CDRs colored as in (A). The CDRs are labeled and colored in order from left to right: H2 (purple), HI (magenta), L3 (yellow), H3 (orange), LI (red), L2 (dark green). (F) Bird’s-eye view of surface representation of the HLA-A*02:01 shown in grey, p53^R175H peptide shown in light green, and the contacting residues colored according to CDRs of the H2-Fab. (G) Schematic representation of (F). (H) Diagram of the orientation angle of the H2-Fab to p53^R175H/HLA-A*02:01. The docking angle of the orientation was calculated from the web server TCR3d which was based on the C^alpha of Cys88 of the disulfide bond of the VL domain and the C^alpha of Cys96 of the disulfide bond of the VH domain of the H2-Fab (red). The arrowed line indicates the direction of orientation with the angle between them.

Figures 71A-71F. Structural basis of H2 specificity and identification of putative cross-reactive peptides. (A) Detailed interactions of the p53^R175H neoantigen with HLA- A*02:01. The peptide (green) and the side chains (grey) of interacting residues of HLA- A*02:01 are represented as sticks. Hydrogen bonds are shown as dashed lines. (B) Perpendicular view of the p53^R175H peptide binding cleft. (C) C-terminus of the peptide (aa Vall73-Cysl76) with Argl74 and Hisl75 surrounded by the interacting residues of CDR-H1 (magenta), -H2 (purple), -H3 (orange) and -L3 (yellow) shown as sticks. Hydrogen bonds are shown as dashed lines. (D) T2 cells were loaded with 10 mM of HMTEVVRHC (SEQ ID NO:l) peptide variants from the positional scanning library and co-incubated with 1 nM H2- scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by cytometric bead array (see Supplementary Materials) and the mean of triplicate wells was used to plot the heatmap. Black boxes represent the parental p53^R175H peptide. (E) Illustration of the binding pattern of H2-scDb as Seq2Logo graph, calculated by dividing the IFN-g value by 10⁴ and using the PSSM-Logo algorithm. (F) T2 cells were loaded with 10 mM of p53^R175H, p53^WT, STAT2, VPS13A, or ZFP3 peptide and co-incubated with 1 nM H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g secretion was measured by ELISA. Data indicate mean ± SD of three technical replicates.

Figures 72A-72B. In vivo antitumor efficacy of H2-scDb. In the early treatment model, NSG mice were engrafted with 1 x 10⁷ human T cells and either 1 x 10⁶ parental KMS26 (A) or 1 x 10⁶ TP53 KO KMS26 (B) on day 0. On day 1, intraperitoneal infusion pumps were placed to administer H2-scDb or isotype control scDb. (C) In the established tumor model, mice were engrafted with 1 x 10⁷ human T cells and 3.5 x 10⁵ parental KMS26 on day 0, followed by administration of H2-scDb or isotype scDb at the specified doses on day 6. Tumor growth was monitored by bioluminescence imaging. N = 4 or 5 mice per group. Color bars denote the radiance (p/sec/cm²/sr) scale at each time point. Plotted data indicate mean ± SD. **P <0.01 and NS denotes no statistical significance compared to isotype control by multiple t-test with Holm-Sidak correction.

Figures 73A-73C. Detection and quantification of p53^R175H neoantigen peptide in cells. (A) COS-7 cells transfected with constructs expressing HLA-A*02:01 and p53^WT or p53^R175H were analyzed for the presentation of the p53^WT HMTEVVRRC (SEQ ID NO: 135) or the p53^R175H HMTEVVRHC (SEQ ID NO: 1) peptide. Isotope labeled peptides were spiked into the assay and served as standards for absolute copy number quantification. Multiple ions (indicated by different colors) fragmented from the target peptide in each sample were measured through mass spectrometer as different SRM transitions and their m/Z values were listed in the figure legend. (B) Expression of p53 protein in COS-7 cells transfected with either the full-length p53^WT or p53^R175H was assessed by Western blotting with anti-p53 antibody (clone DO-1). (C) Cells lines with endogenous HLA-A*02:01 and p53^R175H expression were analyzed for the presentation of p53^R175H neoantigen peptide as described in (A).

Figures 74A-74D. Selection of p53^R175H/HLA-A*02:01 reactive antibodies and their conversion into T cell-retargeting scDb. (A) Flow cytometric screening of phage clones enriched by panning. After 5 rounds of panning, phage clones from the enriched phage pool were isolated by limiting dilution and grown in deep 96-well plates. Supernatants containing individual phage clones were used to assess binding to T2 cells loaded with b2 microglobulin (b2M) only, b2M plus p53^WT peptide (HMTEVVRRC; SEQ ID NO: 135), or b2M plus p53^R175H peptide (HMTEVVRHC; SEQ ID NO:l) via flow cytometry. The median fluorescence intensity (MFI) ratio was defined as MFI (p53^R175 peptide)/MFI (p53^WT peptide). NC, no phage control. (B) Schematic representation of the structure of the T cell- engaging bispecific single-chain diabody (scDb) used in our experiments. VL, variable light domain; VH, variable heavy domain; pHLA, peptide-HLA complex; SL, short linker; LL, long linker. The graph was created with BioRender.com. (C) Screening of scDb clones via IFN-g stimulation by p53-expressing cells. scDbs generated by linking each anti- p53^R175H/HLA-A*02:01 pHLA scFv clone with an anti-CD3 scFv (UCHT1) were co incubated with T cells and COS-7 cells transfected with GFP, HLA-A*02:01 + GFP, HLA- A*02:01 + p53^WT, or HLA-A*02:01 + p53^R175H plasmids at an effectontarget (E:T) ratio of 1:1. After a 20-hr coincubation, the supernatant was harvested for IFN-g detection by ELISA. Arrows indicate clones H2 and H20. A2, HLA-A*02:01. (D) Characterization of H20-scDb. H20-scDb was incubated with biotinylated p53^R175H/HLA-A*02:01 (red) andp53^WT/HLA- A*02:01 (gray) pHLA monomers coated on streptavidin microplates at the specified concentrations, then binding detected with protein L and anti-protein L HRP. Data indicate mean ± SD of three technical replicates.

Figures 75A-75E. Characteristics of scDbs generated by linking H2-scFv with anti- CD3 e scFvs. (A) Expression of scDbs composed of linking H2-scFv with different anti-CD3r scFvs was assessed by anti-6x-His tag Western blotting. (B) Binding of the scDbs to the CD3e/6 heterodimer and CDr was compared using ELISA. (C) The scDbs were co-incubated with T cells and T2 cells pulsed with titrated concentrations of p53^R175H or p53^WT peptide at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Data indicate mean ± SD of three technical replicates. (D) Analytical chromatogram of the purified H2-UCHTl-scDb (H2- scDb) showing absorbance at 280 nm. The retention time of the H2-scDb was marked above the peak. (E) DSF analysis of the negative derivative (RFU vs. temperature) of the H2-scDb. The melting temperature T_m at 69 °C corresponds to the peak/maximum of the first derivative of the curve and the notion of one transition state.

Figures 76A-76F. Reactivity ofH2-scDb against p53^R175H/HLA-A*02:01-expressing tumor cells. (A) TYK-nu and its cisplatin resistant subline TYK-nu.CP-r (75) were cultured with H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Data indicate mean ± SD of six technical replicates and are representative of two independent experiments. (B, C) KMS26 cells were cultured with H2-scDb or an isotype scDb (scFv against an irrelevant pHLA linked with UCHT1 scFv) in the absence or presence of T cells at an E:T ratio of 2:1. IFN-g release was measured by ELISA (B), and cytotoxicity was assessed by luciferase assay (C). Data indicate mean ± SD of three technical replicates. ** <0.01, *** P<0.001 by two tailed t-test. (D-F) H2-scDb-induced polyfunctional T-cell response. T-cell cytotoxicity and cytokine release mediated by H2-scDb in response to (D) KMS26 (cytotoxicity and other effector proteins shown in main text), (E) KLE, and (F) TYK-nu cell line at an E:T ratio of 2: 1 was assessed by antibody-based assays (see Methods). EC50 (M) for each analyte is shown in the corresponding graphs. Data indicate mean ± SD of three technical replicates.

Figures 77A-77B. Flow cytometric evaluation of HLA-A*02 expression. (A) Expression of HLA-A*02 on tumor cell lines was evaluated by flow cytometry. The red histogram represents staining with anti-HLA-A*02 (clone BB7.2), and the gray histogram represents staining with an isotype control. The median fluorescence intensity (MFI) ratio is defined as MFI (anti-HLA-A*02)/MFI (isotype control). (B) HLA-A*02:01 -encoding retrovirus was transduced into cell lines that weakly (AU565, SK-BR-3) or do not detectably (HuCCTl, CCRF-CEM) express HLA-A*02:01. Expression of HLA-A*02:01 in the parental and transduced and sorted cell lines was evaluated by flow cytometry. The red histogram represents staining with anti-HLA-A*02 (clone BB7.2) and the gray histogram represents staining with an isotype control.

Figures 78A-78B. Determination of H2-scDb specificity using CRISPR-edited isogenic cell lines. (A) Expression of p53 protein in the parental and TP53 KO clones of KMS26, KLE, and TYK-nu was assessed by Western blot with anti-p53 antibody (clone DO- 1). (B) Cytotoxicity mediated by H2-scDb in response to parental tumor cell lines and their TP 53 KO counterparts at an E:T ratio of 2: 1 (KMS26, TYK-nu) or 5: 1 (KLE) was measured by the Bio-Glo (KMS-26) or CellTiter-Glo (TYK-nu, KLE) assay. Data indicate mean ± SD of two (TYK-nu) or three (KMS26, KLE) technical replicates and are representative of two independent experiments.

Figures 79A-79D. The H2-Fab-p53^R175H/HLA-A*02:01 complex. (A) H2-scFv was converted into full-length IgG (H2-IgG) and incubated with biotinylated p53^R175H/HLA- A*02:01 (red)andp53^WT/HLA-A*02:01 (gray)pHLA monomers coated on streptavidin microplates at the specified concentrations followed by detection with anti-human IgG HRP. Data indicate mean ± SD of three technical replicates. (B) Illustration depicting the generation of H2-Fab from H2-IgG. (C) Size-exclusion chromatogram of the pHLA-A*02:01 in complex with the H2-Fab. Protein was monitored by A280 nm with a major peak (-100 kDa). (D) Coomassie-stained gradient SDS-PAGE gel of the eluted fractions at 11-17 mL from (C).

Figures 80A-80D. The neoantigen p53^R175H binds to HLA-A*02:01 in a canonical fashion. (A) Bird’s-eye view of the p53^R175H neoantigen interactions with HLA-A*02:01.

The peptide (green) and the side chains (grey) of interacting residues of HLA-A*02:01 are represented as sticks. Hydrogen bonds are shown as dashed lines. The N-terminal His 168 is anchored by three tyrosine residues of HLA-A*02:01, one at the base of the cleft (Tyr7, not shown) and two on a2 (Tyrl59, 171), while its side chain is within hydrogen bonding distance of Lys66 (al) and Thrl63 (a2, not shown). Glu63 of HLA-A*02:01 al forms a hydrogen bond with the backbone amino of Metl69, an anchor residue of p53^R175H that is situated within the hydrophobic B pocket of the HLA. The main chain of Thrl70 is stabilized by a hydrogen bond to Tyr99 (not shown), located at the base of the cleft while the side chain of Glu4 forms a salt-bridge with the side chain of Arg65. Positions 5-8 (Vall72, Vall73, Argl74, Hisl75) of the p53^R175H neoantigen are stabilized by multiple hydrophobic and aliphatic residues with no direct hydrogen bonding contacts to the HLA-A*02:01. Towards the C-terminus of the neoantigen, the carboxyl group of Cysl76, another anchor residue that lies within the F pocket, is secured by Tyr84 (al) and Lysl46 (a2), while the side chain sulfhydryl is near Thrl43 on a2. (B) Surface representation of the HLA-A*02:01 (grey) with the p53^R175H neoantigen shown in green as sticks. Anchor pockets B and F are circled in orange. (C) Structural alignment of the following HLA-A*02:01-bound peptides in the binding pocket: green (this work, PDB ID 6W51), cyan (pl049, PDB ID 2JCC), magenta (NY-ESO-1, PDB ID 3HAE), and light purple (WT1, PDB ID 4WUU). (D) Zoomed in view of (B) with helix al transparent and residues at positions 7 (P7) and 8 (P8) shown as sticks.

Figure 81. Hydrogen bonding pattern of the H2-Fab cage-like configuration. The imidazole ring of Hisl75 was at the center of the cage-like structure. The guanidinium group of Argl74 was at hydrogen bonding distance to the backbone carbonyl of Ala31 (CDR-H1). Another neoantigen-antibody direct contact involves the backbone carbonyl of Vail 73 hydrogen bonding with the side chain of Arg93 (CDR-L3).

Figures 82A-82P. Comparison of the binding orientations among TCRm-pHLA and

TCR-pHLA complexes. (A) Close depiction of binding of the H2-Fab to p53^R175H/HLA- A*02:01 with CDRs colored as in Fig. 4. (B) Binding of a TCR to melanoma-associated antigen 3 (MAGE-A3) and HLA-A*01:01 (PDB ID 5BRZ). Same orientation as (A). The MAGE-A3 TCR displays the canonical, diagonal binding motif to that of most known TCR topologies. (C) Recognition of the 3M4E4 Fab for the NY-ESO-1157-165/HLA- A*02:01complex (PDB ID 3HAE). Same orientation as (A). (D) Binding of the ESK1 Fab to Wilms tumor 1 peptide and HLA-A*02:01 (PDB ID 4WUU). Same orientation as (A). (E, F, G and H) Bird’s-eye view of surface representation of the HLA-A*02:01/*01 :01 colored in grey with the contacting residues of H2-Fab, MAGE- A3 TCR, 3M4E4 Fab, and ESK1 Fab, respectively, colored according to CDRs. (I, J, K and L) Schematic representation of E, F, G, and H, respectively. H2-Fab-p53^R175H/HLA-A*02:01 shows a different mode of antibody recognition compared with other Fab/TCR-pHLA complexes. Schematic representation of Fab/TCR docking angle. The docking angle was calculated from the web server TCR3d which was based on the C^alpha of Cys88 (or equivalent) of the disulfide bond of the VL/O. domain and the C^alpha of Cys96 (or equivalent) of the disulfide bond of the Vi-i/b domain of each antibody and TCR. The arrowed line indicates the direction of orientation with the angle between them.

Figures 83A-83D. Assessment of H2-scDb cross-reactivity. (A) A peptide library was generated by systemically substituting the amino acid at each position of the target peptide (HMTEVVRHC; SEQ ID NO:l) with each of the remaining 19 common amino acids. T2 cells were loaded with each of the variant peptides at 100 mM in the presence of 10 pg/ml b2M and anti-HLA-A*02 antibody (clone BB7.2). HLA-A*02:01 stabilized by peptide binding and was evaluated by flow cytometry (77). Black boxes represent the parental peptide. MFI, median fluorescence intensity. (B) Recognition of the p53^R175H positional scanning library peptides by H2-scDb. T2 cells loaded with variant peptides in the positional scanning library were incubated with InM H2-scDb and T cells at an E:T ratio of 2: 1. IFN-g release was measured by ELISA. Dotted lines represent 20% of parental peptide IFN-g value. The binding motif established by the 20% reactivity cutoff, expressed in PROSITE pattern, was x-[AILMVNQTC]-[ST]-[DE]-[IV]-[IMVST]-R-H-[AILVGHSTYC] (SEQ ID NO: 197). Data indicate mean ± SD of three technical replicates. (C-D) H2-scDb was co-incubated with T cells and COS-7 cells transfected with HLA-A*02:01 and full-length p53^R175H, STAT2, or ZFP3 at an E:T ratio of 5:1. (C) Expression of target proteins by COS-7 cells was assessed by Western blot staining. (D) IFN-g secretion was measured by ELISA. The signals for all of the transfectants except for p53^R175H were indistinguishable and are clustered near the x-axis. Data indicate mean ± SD of three technical replicates and are representative of two independent experiments.

Figures 84A-84D. Assessing in vivo efficacy of scDb in NSG mice. (A) Three days after the injection of KMS26 cells and human T cells, peripheral blood of mice was obtained to assess human T cell engraftment by flow cytometry. Plots shown were gated on live cells. (B) Plasma of mice was collected 3 days before and 3 and 10 days after the implantation of intraperitoneal pumps. Plasma concentration of H2-scDb was measured by ELISA. N = 9 mice. Data shown represent mean ± SEM (C) Serial monitoring of body weight of the mice in the established KMS26 model presented in Fig. 6B. N = 5 mice per group. Data shown represent mean ± SD. (D) To verify the action of H2-scDb is T cell-dependent, NSG mice were engrafted with 5 x 10⁵ parental KMS26 cells on day 0 with and without 1 x 10⁷ human T cells, followed by administration of the specified scDb or vehicle via intraperitoneal pumps on day 6. Tumor growth was monitored by bioluminescence imaging. N = 4 or 5 mice per group. Data shown represent mean ± SD.

DETAILED DESCRIPTION

This document provides methods and materials for assessing a mammal having cancer or suspected of having cancer and/or treating a mammal having cancer. For example, one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can target (e.g., bind to) one or more modified peptides (e.g, peptides present in a peptide-HLA complex such as a peptide-HLA-P2M complex) can be used to assess a mammal having cancer or suspected of having cancer and/or to treat a mammal having a cancer (e.g, a cancer expressing one or more modified peptides). In some cases, one or more molecules includes one or more antigen-binding domains that can bind to a modified peptide can be used to detect the presence or absence of one or more modified peptides in a sample obtained from a mammal having cancer or suspected of having cancer. In some cases, one or more molecules including one or more antigen-binding domains that can bind to a modified peptide can be administered to a mammal having a cancer (e.g, a cancer expressing the modified peptide) to treat the mammal.

As used herein, a modified peptide is a peptide derived from a modified polypeptide. A modified polypeptide can be any appropriate modified polypeptide (e.g, a polypeptide having a disease-causing mutation such as a mutation in an oncogenic or a mutation in a tumor suppressor gene). A modified peptide can have one or more amino acid modifications ( e.g ., substitutions) relative to a WT peptide (e.g, a peptide derived from a WT polypeptide from which the modified polypeptide is derived). A modified peptide also can be referred to as a mutant peptide. In some cases, a modified peptide can be a tumor antigen. Examples of tumor antigens include, without limitation, MANAs, tumor-associated antigens, and tumor- specific antigens. A modified peptide can be any appropriate length. In some cases, a modified peptide can be from about 7 amino acids to about 25 amino acids (e.g, from about 8 amino acids to about 25 amino acids, from about 9 amino acids to about 25 amino acids, from about 10 amino acids to about 25 amino acids, from about 11 amino acids to about 25 amino acids, from about 12 amino acids to about 25 amino acids, from about 13 amino acids to about 25 amino acids, from about 15 amino acids to about 25 amino acids, from about 18 amino acids to about 25 amino acids, from about 20 amino acids to about 25 amino acids, from about 7 amino acids to about 22 amino acids, from about 7 amino acids to about 20 amino acids, from about 7 amino acids to about 18 amino acids, from about 7 amino acids to about 15 amino acids, from about 7 amino acids to about 12 amino acids, from about 7 amino acids to about 10 amino acids, from about 7 amino acids to about 9 amino acids, from about 8 amino acids to about 22 amino acids, from about 10 amino acids to about 18 amino acids, from about 12 amino acids to about 15 amino acids, from about 8 amino acids to about 12 amino acids, from about 12 amino acids to about 18 amino acids, from about 18 amino acids to about 22 amino acids, or from about 9 amino acids to about 10 amino acids) in length. For example, a modified peptide can be about 9 amino acids in length. For example, a modified peptide can be about 10 amino acids in length. A modified peptide can be derived from any modified polypeptide. Examples of modified polypeptides from which modified peptides described herein can be derived include, without limitation, p53 and RAS (e.g, KRAS, HRAS, and NRAS). A modified peptide can include any appropriate modification. In some cases, modified peptides described herein can include one or more modifications (e.g, mutations) shown in Table 1. Table 1. Modified peptides.

A modified peptide described herein ( e.g ., a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can be in a complex with any appropriate HLA. An HLA can be any appropriate HLA allele. In some cases, an HLA can be a class I HLA (e.g., HLA- A, HLA-B, and HLA-C) allele. In some cases, an HLA can be a class II HLA (e.g., HLA-DP, HLA-DM, HLA-DOA, HLA-DOB, HLA-DQ, and HLA-DR) allele. Examples of HLA alleles that a modified peptide described herein can complex with include, without limitation, HLA-Al and HLA- A2. Exemplary HLA alleles for particular modified peptides are shown in Table 1. For example, a modified peptide derived from a modified p53 polypeptide (e.g, HMTEVVRHC (SEQ ID NO: 1)) can be in a complex with HLA-A2 and b2M. For example a modified peptide derived from a modified H/K/N RAS polypeptide (e.g, ILDTAGHEEY (SEQ ID NO:2), ILDTAGLEEY (SEQ ID NO:3), and ILDTAGREEY (SEQ ID NO:4)) can be in a complex with HLA-Al (e.g, can be in a complex with HLA-Al and b2M).

This document provides molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4). In some cases, a molecule including one or more antigen- binding domains that can bind to a modified peptide described herein does not target (e.g, does not bind to) an uncomplexed modified peptide described herein (e.g., a modified peptide described herein that is not present in a complex (e.g, a peptide-HLA^2M complex)). In some cases, a molecule including one or more antigen-binding domains that can bind to a modified peptide described herein does not target (e.g, does not bind to) a WT peptide (e.g, a peptide derived from a WT polypeptide from which the modified polypeptide is derived). A molecule including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide described herein can be any appropriate type of molecule. In some cases, a molecule can be a monovalent molecule (e.g., containing a single antigen-binding domain). In some cases, a molecule can be a multivalent molecule (e.g, containing two or more antigen-binding domains and simultaneously targeting two or more antigens). For example, a bispecific molecule can include two antigen-binding domains, a trispecific molecule can include three antigen-binding domains, a quadruspecific molecule can include four antigen-binding domains, etc. Examples of molecules that contain antigen-binding domains include, without limitation, antibodies, antibody fragments, scFvs, chimeric antigen receptors (CARs), T cell receptors (TCRs), TCR mimics, tandem scFvs, bispecific T cell engagers, diabodies, scDbs, scFv-Fcs, bispecific antibodies, bispecific single-chain Fes, dual- affinity re-targeting antibodies (DARTs), and any other molecule that includes at least one variable heavy chain (VH) and at least one variable light chain (VL). Any of these molecules can be used in accordance with materials and methods described herein. In some cases, an antigen-binding domain can be a scFv. For example, a molecule including one or more antigen-binding domains (e.g, one or more scFvs) that can bind to a modified peptide described herein can be a CAR. For example, a molecule including two scFvs that can bind to a modified peptide described herein can be a single-chain diabody (scDb).

In some cases, when a molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) is a multivalent molecule (e.g, a bispecific molecule), a first antigen-binding domain can bind to a modified peptide described herein and a second antigen-binding domain can bind to an effector cell (e.g. , an antigen present on an effector cell). Examples of effector cells include, without limitation, T cells, natural killer (NK) cells, natural killer T (NKT) cells, B cells, plasma cells, macrophages, monocytes, microglia, dendritic cells, neutrophils, fibroblasts, and mast cells. Examples of antigens present on effector cells include, without limitation, CD3, CD4, CD8, CD28, NKG2D, PD-1, CTLA-4, 4-1BB, 0X40, ICOS, CD27, Fc receptors (e.g., CD16a), and any other effector cell surface receptors. In some cases, a molecule described herein can include a first antigen-binding domain that can bind to a modified peptide described herein and a second antigen-binding domain that can bind to an antigen present on a T cell (e.g, CD3). In some cases, sequences (e.g., scFv sequences) that can bind to CD3 can be as shown in Table 4. In some cases, sequences (e.g, scFv sequences) that can bind to CD3 can be as described elsewhere (see, e.g., Rodrigues etal, 1992 Int J Cancer Suppl . 7:45-50; Shalaby etal, 1992 J Exp Med. 175:217-25; Brischwein etal, 2006 Mol Immunol . 43:1129-43; Li etal. , 2005 Immunology. 116:487-98; WO2012162067; US20070065437; US20070065437; US20070065437; US20070065437; US20070065437; and US20070065437). In some cases, a molecule described herein can include a first antigen-binding domain that can bind to a modified peptide described herein and a second antigen-binding domain that can bind to an antigen present on a NK cell (e.g, CD 16a or NKG2D). In some cases, sequences (e.g, scFv sequences) that can bind to CD 16a can be as shown in Table 5. By binding both the modified peptide and the effector cell, the multivalent molecule can bring the cell expressing the modified peptide (e.g, as part of the HLA complex) into proximity with the effector cell, permitting the effector cell to act on the cell expressing the modified peptide.

In some cases, when a molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) is a multivalent molecule (e.g, a bispecific molecule), a molecule can be in any appropriate format which includes at least one VH and at least one VL. For example, a VH and a VL can be in any appropriate orientation. In some cases, a VH can be N-terminal to the VL. In some cases, a VH can be C-terminal to the VL. In some cases, a linker amino acid sequence can be positioned between the VH and VL.

In some cases, when a bispecific molecule includes tandem scFvs, the tandem scFvs can be in any appropriate orientation. Examples of tandem scFv orientations including scFv- A and scFv-B include, without limitation, VLA-LL-VHA-SL-VLB-LL-VHB, VLA-LL- VHA-SL-VHB-LL-VLB, VHA-LL-VLA-SL-VLB-LL-VHB, VHA-LL-VLA-SL-VHB-LL- VLB, VLB -LL- VHB - SL- VL A-LL- VH A, VLB-LL-VHB-SL-VHA-LL-VLA, VHB-LL- VLB-SL-VLA-LL-VHA, and VHB-LL-VLB-SL-VHA-LL-VLA, where SL is a short linker and LL is a long linker. A short linker can be from about 3 amino acids to about 10 amino acids in length. A short linker can include any appropriate amino acids (e.g, glycines and serines) in any appropriate combination. A long linker can be from about 10 amino acids to about 25 amino acids in length. A long linker can include any appropriate amino acids (e.g., glycines and serines) in any appropriate combination. In some cases, when a bispecific molecule is a diabody, the diabody can be in any appropriate orientation. Examples of diabody orientations including scFv-A and scFv-B include, without limitation, VLA-SL-VHB and VLB-SL-VHA, VLA-SL-VLB and VHB-SL-

VHA, VHA-SL-VLB and VHB-SL-VLA, VLB-SL-VHA and VLA-SL-VHB, VLB-SL-VLA and VHA-SL-VHB, and VHB-SL-VLA and VHA-SL-VLB, where SL is a short linker. A short linker can be from about 3 amino acids to about 10 amino acids in length. A short linker can include any appropriate amino acids ( e.g ., glycines and serines) in any appropriate combination.

In some cases, when a bispecific molecule is a scDb, the scDb can be in any appropriate orientation. Examples of scDb orientations including scFv-A and scFv-B include, without limitation, VLA-SL-VHB-LL-VLB-SL-VHA, VHA-SL-VLB-LL-VHB-SL- VLA, VLA-SL-VLB-LL-VHB-SL-VHA, VHA-SL-VHB-LL- VLB-SL-VLA, VLB-SL- VHA-LL-VLA-SL-VHB, VHB-SL-VLA-LL-VHA-SL-VLB, VLB-SL-VLA-LL-VHA-SL-

VHB, and VHB-SL-VHA-LL-VLA-SL-VLB, where SL is a short linker and LL is a long linker. A short linker can be from about 3 amino acids to about 10 amino acids in length. A short linker can include any appropriate amino acids (e.g., glycines and serines) in any appropriate combination. A long linker can be from about 10 amino acids to about 25 amino acids in length. A long linker can include any appropriate amino acids (e.g, glycines and serines) in any appropriate combination.

In some cases, when a bispecific molecule is a scFv-Fc, the scFv-Fc can be in any appropriate orientation. Examples of scFv-Fc orientations including scFv-Fc-A, scFv-Fc-B, and an Fc domain include, without limitation, VLA-LL-VHA-hinge-Fc and VLB-LL-VHB- hinge-Fc, VHA-LL-VLA-hinge-Fc and VHB-LL-VLB-hinge-Fc, VLA-LL-VHA-hinge-Fc and VHB-LL-VLB-hinge-Fc, VHA-LL-VLA-hinge-Fc and VLB-LL-VHB-hinge-Fc, where LL is a long linker. A long linker can be from about 10 amino acids to about 25 amino acids in length. A long linker can include any appropriate amino acids (e.g, glycines and serines) in any appropriate combination. In some cases, an Fc domain in a scFv-Fc can include one or more modifications to increase heterodimerization and/or to decrease homodimerization of the scFv-Fc. In some cases, an Fc domain in a scFv-Fc can exclude a hinge domain. In some cases, an Fc domain in a scFv-Fc can be at the N-terminus of the scFv.

In some cases, when a bispecific molecule is a bispecific single-chain Fc, the bispecific single-chain Fc can be in any appropriate orientation. Examples of bispecific single-chain Fc orientations include, without limitation, VLA-LL-VHA-SL-VHB-LL-VLB- SL-hinge-CH2-CH3 -LL-hinge-CH2-CH3 , VLA-LL-VHA-SL-VLB-LL-VHB-SL-hinge- CH2-CH3 -LL-hinge-CH2-CH3 , VHA-LL- VL A- SL- VLB -LL- VHB- SL-hinge-CH2-CH3 -LL- hinge-CH2-CH3 , VHA-LL- VLA-SL-VHB-LL-VLB-SL-hinge-CH2-CH3-LL-hinge-CH2- CH3, and VLA-SL-VHB-LL-VLB-VHA-SL-hinge-CH2-CH3-LL-hinge-CH2-CH3, where SL is a short linker and LL is a long linker. A short linker can be from about 3 amino acids to about 8 amino acids in length. A short linker can include any appropriate amino acids ( e.g ., glycines and serines) in any appropriate combination. A long linker can be from about 10 amino acids to about 25 amino acids in length. A long linker can include any appropriate amino acids (e.g., glycines and serines) in any appropriate combination. Any appropriate Fc domain can be used in a bispecific single-chain Fc. In some cases, an Fc domain can include an amino acid sequence derived from an IgG (e.g., a natural IgG). In some cases, an Fc domain can include an amino acid sequence that includes one or more modifications (e.g., one or more modifications to increase stability of the molecule and/or to increase or decrease binding to one or more Fc receptors). In some cases, an Fc domain that can be used in a bispecific single-chain Fc can exclude a hinge domain. In some cases, an Fc domain that can be used in a bispecific single-chain Fc can be at the N-terminus of the scFvs. In some cases, an Fc domain that can be used in a bispecific single-chain Fc can be as described elsewhere (see, e.g., International Patent Application Publication No. WO 2017/134134 A1 at, for example, SEQ ID NOs:25-32; and International Patent Application Publication No. WO 2017/134158 A1 at, for example, Table 38; and SEQ ID NOs:25-32).

A molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can include any appropriate complementarity determining regions (CDRs). For example, a molecule including one or more antigen-binding domains that can bind to a modified peptide described herein can include a variable heavy chain (VH) having three VH complementarity determining regions (CDR-VHs) and a variable light chain (VL) having three VL CDRs (CDR-VLs). For example, a molecule that can bind to a modified peptide derived from a modified p53 polypeptide (e.g, HMTEVVRHC (SEQ ID NO:l)) can include one of each of the CDRs set forth below: CDR-VL1: QDVNTA (SEQ ID NO: 5);

CDR-VL2: SAS and SAY;

CDR-VL3: QQYSRYSPVTF (SEQ ID NO: 6), QQQSSTPVTF (SEQ ID NO: 7), QQSSYYPNTF (SEQ ID NO: 8), QQQWSSPDTF (SEQ ID NO: 9), QQSNAYPITF (SEQ ID NO: 10);

CDR-VH1: GFNVYASGM (SEQ ID NO: 11), GFNVYQSDM (SEQ ID NO: 12), GFNLYQRDM (SEQ ID NO: 13), GFNLSYYDM (SEQ ID NO: 14), GFNLNSYYM (SEQ ID NO: 15);

CDR-VH2: KIYPDSDYTY (SEQ ID NO: 16), TIWPYSGYTY (SEQ ID NO: 17), GLLYGSDHTE (SEQ ID NO: 18), LIYYGSGYTY (SEQ ID NO: 19), MIIPGYGYTN (SEQ ID NO: 20); and

CDR-VH3: SRD S SF YYVYAMD Y (SEQ ID NO: 21), SRDGMYAFDY (SEQ ID NO: 22), SRATYEEAFDY (SEQ ID NO: 23), SRGS YVSGMD Y (SEQ ID NO: 24), SRSYYMYMDY (SEQ ID NO: 25).

For example, a molecule that can bind to a modified peptide derived from a modified H/K/N RAS polypeptide Q61H (e.g., ILDTAGHEEY (SEQ ID NO:2)) can include one of each of the CDRs set forth below:

CDR-VL1: QDVNTA (SEQ ID NO: 5);

CDR-VL2: SAS;

CDR-VL3: QQVIYYPFTF (SEQ ID NO: 26), QQYDYYPFTF (SEQ ID NO: 27), QQSIYYPFTF (SEQ ID NO: 28), QQSSYSPWTF (SEQ ID NO: 184), QQSFSTPITF (SEQ ID NO: 29), QQGEYSPLTF (SEQ ID NO: 30), QQTYYTPVTF, (SEQ ID NO: 31);

CDR-VH1: GFNLYSYAI (SEQ ID NO: 32), GFNISYEAM (SEQ ID NO: 33), GFNLYTSQM (SEQ ID NO: 34), GFNVFGYAI (SEQ ID NO: 35), GFNISPWDM (SEQ ID NO: 36), GFNISEYLM (SEQ ID NO: 37), GFNVFESAM (SEQ ID NO: 38), GFNISHYVM (SEQ ID NO: 39);

CDR-VH2: LLYPDYGVTS (SEQ ID NO: 40), LIYPNHGITS (SEQ ID NO: 41), LVYPGYYVTS (SEQ ID NO: 42), EVYPGYD VT S (SEQ ID NO: 43), QLYPSSGYTN (SEQ ID NO: 44), LLPPGLSYTN (SEQ ID NO: 45), WVYGSYDYTY (SEQ ID NO: 46), DFYPHSDSTY (SEQ ID NO: 47); and CDR-VH3: SRYRSYEYSVSSYSYSAMDY (SEQ ID NO: 48), SRYSSSAMDY (SEQ ID NO: 49), SRGAYYY S S AMD Y (SEQ ID NO: 50), SRY S WAGAFD Y (SEQ ID NO: 51), SRSVYWSLDY (SEQ ID NO: 52), SRY GYYAFD Y (SEQ ID NO: 53), SRSFAYFQAMDY (SEQ ID NO: 54), SRYQSYSFDY (SEQ ID NO: 55).

For example, a molecule that can bind to a modified peptide derived from a modified H/K/N RAS polypeptide Q61L (e.g, ILDTAGLEEY (SEQ ID NO:3) can include one of each of the CDRs set forth below:

CDR-VL1: QDVNTA (SEQ ID NO: 5);

CDR-VL2: SAS;

CDR-VL3: QQASRQPYTF (SEQ ID NO: 56), QQAVSYPWTF (SEQ ID NO: 57), QQTSSYPITF (SEQ ID NO: 58), QQSWYSPSTF (SEQ ID NO: 59), QQSYYAPITF (SEQ ID NO: 60), QQSYYSPWTF (SEQ ID NO: 61), QQAYYPPWTF (SEQ ID NO: 62), QQSYSSGPVTF (SEQ ID NO: 63), QQTYYYPFTF (SEQ ID NO: 64), QQSYYPYYPWTF (SEQ ID NO: 65), QQYDRPITF (SEQ ID NO: 66);

CDR-VH1: GFNFSESGM (SEQ ID NO: 67), GFNISSSGI (SEQ ID NO: 68), GFNIYWYGM (SEQ ID NO: 69), GFNISASGM (SEQ ID NO: 70), GFNFSYYGM (SEQ ID NO: 71), GFNISYSNI (SEQ ID NO: 72), GFNVSRWAM (SEQ ID NO: 73), GFNFSYGGI (SEQ ID NO: 74), GFNLYAWGM (SEQ ID NO: 75), GFNVSHSAM (SEQ ID NO: 76), GFNIYYEAM (SEQ ID NO: 77)

CDR-VH2: HFSGDSGYTY (SEQ ID NO: 78), MVYGGSGYTN (SEQ ID NO: 79), QVYPWSGFTY (SEQ ID NO: 80), WIWGGSSYTY (SEQ ID NO: 81), WIYPFSGYTN, (SEQ ID NO: 82), MIYGTRGGTY (SEQ ID NO: 83), RVYPSGYLTY (SEQ ID NO: 84), MIYPLTGYTN (SEQ ID NO: 85), LVYGGWGSTS (SEQ ID NO: 86), TVHPDWGNTY (SEQ ID NO: 87), QIYPWNDYTY (SEQ ID NO: 88); and

CDR-VH3: SRYMYYSGYFD Y (SEQ ID NO: 89), SRWAHY S AYMD Y (SEQ ID NO: 90), SRDYYSYSLDY (SEQ ID NO: 91), SRGQYLSYMDY (SEQ ID NO: 92), SREYYSRAFDY (SEQ ID NO: 93), SRYYSYAMDY (SEQ ID NO: 94), SRNMQSYMDY (SEQ ID NO: 95), SRDYYYSVDV (SEQ ID NO: 96), SRAGS SKMS AGAFD Y (SEQ ID NO: 97), SRWQQYYYSFDY (SEQ ID NO: 98), SRNYYAATMDY (SEQ ID NO: 99) For example, a molecule that can bind to a modified peptide derived from a modified H/K/N RAS polypeptide Q61R ( e.g ., ILDTAGREEY (SEQ ID NO:4) can include one of each of the CDRs set forth below:

CDR-VL1: QDVNTA (SEQ ID NO: 5); CDR-VL2: SAS;

CDR-VL3: QQSYTSPLTF (SEQ ID NO: 100), QQYWYYYPITF (SEQ ID NO: 101), QQSYYAPITF (SEQ ID NO: 60), QQYYLYQPITF (SEQ ID NO: 102), QQYSNYPLTF (SEQ ID NO: 103), QQYASDPITF (SEQ ID NO: 104), QQYSYDPITF (SEQ ID NO: 105), QQYIYDPVTF (SEQ ID NO: 106), QQLMYDPITF (SEQ ID NO: 107); CDR-VH1: GFNIYYGVM (SEQ ID NO: 108), GFNIYSYDM (SEQ ID NO: 109),

GFNVQWSHM (SEQ ID NO: 110), GFNIGMYTM (SEQ ID NO: 111), GFNVFYGSM (SEQ ID NO: 112), GFNLDYGWM (SEQ ID NO: 113), GFNFSYSAM (SEQ ID NO: 114), GFNVDWAWM (SEQ ID NO: 115), GFNFGTYWM, (SEQ ID NO: 116);

CDR-VH2: MIYPDSSWTY (SEQ ID NO: 117), ISPGGSYTY (SEQ ID NO: 118), RLSPPSGYTN (SEQ ID NO: 119), LVYPD SGYTN (SEQ ID NO: 120), FIGPDSTYTY (SEQ ID NO: 121), WVVPGSDYTD (SEQ ID NO: 122), DVVPDGDWTY (SEQ ID NO: 123), WVVGGSDYTY (SEQ ID NO: 124), WFLPDYDYTL (SEQ ID NO: 125); and

CDR-VH3: SRDQDFHYMNYYLSYALDY (SEQ ID NO: 126), SRSAFTGYFDV (SEQ ID NO: 127), SRLILSKGGYGWAMDY (SEQ ID NO: 128), SRYTWQSMDY (SEQ ID NO: 129), SRDLGSAYAMDY (SEQ ID NO: 130), SRFHYTAFDV (SEQ ID NO: 131), SRGWYALDY (SEQ ID NO: 132), SRSYYYAFDY (SEQ ID NO: 133), SRHGEYAFDY (SEQ ID NO: 134).

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 2. MANAbody complementarity-determining region (CDR) sequences of light (L) chains and heavy (H) chains.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

In some cases, a molecule including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide described herein (e.g., a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can include any appropriate set of CDR sequences (e.g, any of the CDR sequence sets described herein).

A molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can include any appropriate sequence. For example, a molecule that can bind to a modified peptide derived from a modified p53 polypeptide (e.g, HMTEVVRHC (SEQ ID NO: 1)) can include, without limitation, the scFv sequence set forth in any one of SEQ ID NO: 137, SEQ ID NO: 138, SEQ ID NO: 139, SEQ ID NO: 140, or SEQ ID NO: 141. For example, a molecule that can bind to a modified peptide derived from a modified H/K/N RAS polypeptide (e.g, ILDTAGHEEY (SEQ ID NO:2), ILDTAGLEEY (SEQ ID NO:3), or ILDTAGREEY (SEQ ID NO:4)) can include, without limitation, the scFv sequence set forth in any one of SEQ ID NO: 142, SEQ ID NO: 143, SEQ ID NO: 144, SEQ ID NO: 145, SEQ ID NO: 146, SEQ ID NO: 147, SEQ ID NO: 148, SEQ ID NO: 149, SEQ ID NO: 150, SEQ ID NO:151, SEQ ID NO: 152, SEQ ID NO:153, SEQ ID NO:154, SEQ ID NO: 155, SEQ ID NO: 156, SEQ ID NO: 157, SEQ ID NO: 158, SEQ ID NO: 159, SEQ ID NO: 160, SEQ ID NO: 161, SEQ ID NO: 162, SEQ ID NO: 163, SEQ ID NO: 164, SEQ ID NO: 165, SEQ ID NO: 166, SEQ ID NO: 167, SEQ ID NO: 168, or SEQ ID NO: 169. Examples of sequences (e.g, scFv sequences) that can bind to particular modified peptides are shown in Table 3 and Table 12. In some cases, a molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can have a sequence that deviates from a sequence shown in Table 3 and Table 12, sometimes referred to as a variant sequence. For example, a molecule including one or more antigen-binding domains that can bind to a modified peptide described herein can have at least 75% sequence identity (e.g, at least 80% sequence identity, at least 85% sequence identity, at least 90% sequence identity, at least 95% sequence identity, at least 96% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or more) to any of the sequences shown in Table 3 and Table 12, provided the variant sequence maintains the ability to bind to a modified peptide described herein. For example, a molecule including one or more antigen-binding domains that can bind to a modified peptide described herein can have one or more (e.g., one, two, three, four, five, six, seven, eight, nine, ten, or more) modifications (e.g., one or more amino acid substitutions) as compared to the sequences shown in Table 3 and Table 12, provided the variant sequence maintains the ability to bind to a modified peptide described herein. In some cases, a molecule including one or more antigen-binding domains that can bind to a modified peptide described herein can include any appropriate set of CDR sequences described herein, and any sequence deviations from a sequence shown in Table 3 and Table 12 can be in the scaffold sequence(s).

A molecule including one or more antigen-binding domains (e.g., scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can be attached (e.g, covalently or non-covalently attached) to a label (e.g, a detectable label). A detectable label can be any appropriate label. In some cases, a label can be used to assist in detecting the presence or absence of one or more modified peptides described herein. For example, a molecule described herein that is labelled can be used in vitro to detect cancer cells (e.g, cancer cells expressing a modified peptide described herein) in a sample obtained from a mammal. In some cases, a label (e.g, a detectable label) can be used to assist in determining the location of one or more modified peptides described herein. For example, molecule described herein that is labelled can be used in vivo to monitor anti tumor therapy and/or to detect cancer cells (e.g, cancer cells expressing a modified peptide described herein) in a mammal. Examples of labels that can be attached to a molecule described herein include, without limitation, radionuclides, contrast agents used in magnetic resonance imaging (MRI), computed tomography (CT), ultrasound (US), and other imaging modalities, chromophores, enzymes, and fluorescent molecules (e.g, green fluorescent protein and near-IR fluorescence).

A molecule including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, and SEQ ID NO:4) can be attached (e.g., covalently or non-covalently attached) to a therapeutic agent. A therapeutic agent can be any therapeutic agent. In some cases, a therapeutic agent can be an anti-cancer agent. Examples of therapeutic agents that can be attached to a molecule described herein include, without limitation, anti-cancer agents such as monomethyl auri statin E (MMAE), monomethyl auri statin F (MMAF), maytansine, mertansine/emtansine (DM1), ravtansine/soravtansine (DM4), SN-38, calicheamicin, D6.5, dimeric pyrrol obenzodiazepines (PBDs), a-amantin (AAMT), PNU- 159682, ricin, pseudomonas exotoxin A, diphtheria toxin, and gelonin.

This document also provides methods for using one or more molecules including one or more antigen-binding domains (e.g., scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4). For example, one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can target (e.g, bind to) one or more modified peptides can be used to assess a mammal having cancer or suspected of having cancer and/or to treat a mammal having a cancer (e.g, a cancer expressing one or more modified peptides such as a p53 R175H MANA, a RAS Q61H/L/R MANA, and/or a RAS G12V MANA). In some cases, one or more molecules includes one or more antigen-binding domains that can bind to a modified peptide can be used to detect the presence or absence of one or more modified peptides in a sample obtained from a mammal having cancer or suspected of having cancer. In some cases, one or more molecules including one or more antigen-binding domains that can bind to a modified peptide can be administered to a mammal having a cancer (e.g, a cancer expressing the modified peptide) to treat the mammal. Administration of one or more molecules including one or more antigen binding domains that can bind to a modified peptide described herein to a mammal (e.g, human) having a cancer can be effective to treat the mammal.

Any type of mammal can be assessed and/or treated as described herein. Examples of mammals that can be assessed and/or treated as described herein include, without limitation, primates (e.g, humans and non-human primates such as chimpanzees, baboons, or monkeys), dogs, cats, pigs, sheep, rabbits, mice, and rats. In some cases, a mammal can be a human.

A mammal can be assessed and/or treated for any appropriate cancer. In some cases, a cancer can express one or more modified peptides (e.g, one or more MANAs) described herein. A cancer can be a primary cancer. A cancer can be a metastatic cancer. A cancer can include one or more solid tumors. A cancer can include one or more non-solid tumors. Examples of cancers that can be assessed as described herein ( e.g ., based at least in part on the presence of one or more modified peptides described herein) and/or treated as described herein (e.g, by administering one or more molecules including one or more antigen-binding domains (e.g., scFvs) that can bind to a modified peptide described herein) include, without limitation, blood cancers (e.g, Hodgkin’s lymphoma, non-Hodgkin’s lymphoma, acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), multiple myeloma, MDS, and myeloproliferative diseases), lung cancers, pancreatic cancers, gastric cancers, colon cancers (e.g, colorectal cancers), ovarian cancers, endometrial cancers, biliary tract cancers, liver cancers, bone and soft tissue cancers (e.g., sarcomas), breast cancers, prostate cancers, esophageal cancers, stomach cancers, kidney cancers, head and neck cancers, brain cancers (e.g, glioblastoma multiforme and astrocytomas), thyroid cancers, germ cell tumors, and melanomas.

When assessing a mammal having cancer or suspected of having cancer, one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be used to assess for the presence or absence of one or more modified peptides described herein. For example, the presence, absence, or level of one or more modified peptides described herein in a sample obtained from a human can be used to determine whether or not the human has a cancer. In some cases, the presence of one or more modified peptides described herein in a sample obtained from a mammal can be used to identify the mammal as having a cancer. For example, a mammal can be identified as having a cancer when a sample obtained from the mammal has one or more modified peptides described herein.

Any appropriate sample obtained from a mammal can be assessed for the presence, absence, or level of one or more modified peptides described herein. For example, biological samples such as tissue samples (e.g, breast tissue, and cervical tissue such as from a Papanicolaou (Pap) test), fluid samples (e.g, blood, serum, plasma, urine, saliva, sputum, and cerebrospinal fluid), and solid samples (e.g. stool) can be obtained from a mammal and assessed for the presence, absence, or level of one or more modified peptides described herein. Any appropriate method can be used to detect the presence, absence, or level of one or more modified peptides described herein. For example, sequencing techniques including, but not limited to, Sanger sequencing, chemical sequencing, nanopore sequencing, sequencing by ligation (SOLiD sequencing), sequencing with mass spectrometry, whole exome sequencing, whole genome sequencing, and/or next-generation sequencing can be used to determine the presence, absence, or level of one or more modified peptides described herein in a sample obtained from a mammal.

When treating a mammal having cancer, one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide described herein (e.g., a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be administered to a mammal having cancer to treat the mammal. In some cases, a mammal can have a cancer expressing one or more modified peptides described herein. For example, one or more molecules including one or more antigen-binding domains that can bind to a modified peptide described herein can be administered to a mammal having a cancer expressing that modified peptide to treat the mammal. For example, one or more molecules including one or more scFvs that can bind to a modified peptide described herein (e.g, one or more scDbs) can be administered to a mammal having a cancer expressing that modified peptide to treat the mammal.

In some cases, one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be administered to a mammal (e.g, a mammal having a cancer) once or multiple times over a period of time ranging from days to weeks.

In some cases, one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be formulated into a composition (e.g, a pharmaceutically acceptable composition) for administration to a mammal (e.g, a mammal having a cancer). For example, one or more antigen-binding domains that can bind to a modified peptide described herein can be formulated together with one or more pharmaceutically acceptable carriers (additives), excipients, and/or diluents. In some cases, a pharmaceutically acceptable carrier, excipient, or diluent can be a naturally occurring pharmaceutically acceptable carrier, excipient, or diluent. In some cases, a pharmaceutically acceptable carrier, excipient, or diluent can be a non-naturally occurring (e.g, an artificial or synthetic) pharmaceutically acceptable carrier, excipient, or diluent. Examples of pharmaceutically acceptable carriers, excipients, and diluents that can be used in a composition described herein include, without limitation, sucrose, lactose, starch (e.g, starch glycolate), cellulose, cellulose derivatives (e.g, modified celluloses such as microcrystalline cellulose and cellulose ethers like hydroxypropyl cellulose (HPC) and cellulose ether hydroxypropyl methylcellulose (HPMC)), xylitol, sorbitol, mannitol, gelatin, polymers (e.g, polyvinylpyrrolidone (PVP), polyethylene glycol (PEG), crosslinked polyvinylpyrrolidone (crospovidone), carboxymethyl cellulose, polyethylene-polyoxypropylene-block polymers, and crosslinked sodium carboxymethyl cellulose (croscarmellose sodium)), titanium oxide, azo dyes, silica gel, fumed silica, talc, magnesium carbonate, vegetable stearin, magnesium stearate, aluminum stearate, stearic acid, antioxidants (e.g, vitamin A, vitamin E, vitamin C, retinyl palmitate, and selenium), citric acid, sodium citrate, benzyl alcohol, lysine hydrochloride, trehalose dihydrate, sodium hydroxide, parabens (e.g, methyl paraben and propyl paraben), petrolatum, dimethyl sulfoxide, mineral oil, serum proteins (e.g., human serum albumin), glycine, sorbic acid, potassium sorbate, water, salts or electrolytes (e.g, saline, protamine sulfate, di sodium hydrogen phosphate, potassium hydrogen phosphate, sodium chloride, and zinc salts), colloidal silica, magnesium trisilicate, polyacrylates, waxes, wool fat, lecithin, and corn oil. In some cases, a pharmaceutically acceptable carrier, excipient, or diluent can be an antiadherent, a binder, a colorant, a disintegrant, a flavor (e.g, a natural flavor such as a fruit extract or an artificial flavor), a glidant, a lubricant, a preservative, a sorbent, and/or a sweetener.

A composition (e.g, a pharmaceutical composition) containing one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be formulated into any appropriate dosage form. Examples of dosage forms include solid or liquid forms including, without limitation, gums, capsules, tablets (e.g, chewable tablets, and enteric coated tablets), suppositories, liquids, enemas, suspensions, solutions ( e.g ., sterile solutions), sustained-release formulations, delayed-release formulations, pills, powders, and granules.

A composition containing one or more molecules including one or more antigen binding domains (e.g., scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be designed for oral, parenteral (including subcutaneous, intramuscular, intravenous, and intradermal), or intratumoral administration. Compositions suitable for parenteral administration include aqueous and non-aqueous sterile injection solutions that can contain anti-oxidants, buffers, bacteriostats, and solutes that render the formulation isotonic with the blood of the intended recipient. The formulations can be presented in unit-dose or multi-dose containers, for example, sealed ampules and vials, and may be stored in a freeze dried (lyophilized) condition requiring only the addition of the sterile liquid carrier, for example, water for injections, immediately prior to use. Extemporaneous injection solutions and suspensions may be prepared from sterile powders, granules, and tablets.

A composition containing one or more molecules including one or more antigen binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be administered using any appropriate technique and to any appropriate location. A composition including one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can be administered locally (e.g, intratumorally) or systemically. For example, a composition provided herein can be administered locally by intratumoral administration (e.g, injection into tumors) or by administration into biological spaces infiltrated by tumors (e.g. intraspinal administration, intracerebellar administration, intraperitoneal administration and/or pleural administration). For example, a composition provided herein can be administered systemically by oral administration or by intravenous administration (e.g, injection or infusion) to a mammal (e.g, a human).

Effective doses can vary depending on the risk and/or the severity of the cancer, the route of administration, the age and general health condition of the subject, excipient usage, the possibility of co-usage with other therapeutic treatments such as use of other agents, and the judgment of the treating physician. An effective amount of a composition containing one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide described herein (e.g., a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO:l, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be any amount that treats a cancer present within the subject without producing significant toxicity to the subject. If a particular subject fails to respond to a particular amount, then the amount of one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can be increased (e.g, by two-fold, three-fold, four-fold, or more). After receiving this higher amount, the mammal can be monitored for both responsiveness to the treatment and toxicity symptoms, and adjustments made accordingly. The effective amount can remain constant or can be adjusted as a sliding scale or variable dose depending on the subject’s response to treatment. Various factors can influence the actual effective amount used for a particular application.

For example, the frequency of administration, duration of treatment, use of multiple treatment agents, route of administration, and severity of the condition (e.g, cancer) may require an increase or decrease in the actual effective amount administered.

The frequency of administration of one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be any frequency that effectively treats a mammal having a cancer without producing significant toxicity to the mammal. For example, the frequency of administration of one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can be from about two to about three times a week to about two to about three times a year.

In some cases, a subject having cancer can receive a single administration of one or more antibodies described herein. The frequency of administration of one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can remain constant or can be variable during the duration of treatment. A course of treatment with a composition containing one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can include rest periods. For example, a composition containing one or more molecules including one or more antigen-binding domains ( e.g ., scFvs) that can bind to a modified peptide described herein can be administered every other month over a two-year period followed by a six-month rest period, and such a regimen can be repeated multiple times. As with the effective amount, various factors can influence the actual frequency of administration used for a particular application. For example, the effective amount, duration of treatment, use of multiple treatment agents, route of administration, and severity of the condition (e.g., cancer) may require an increase or decrease in administration frequency.

An effective duration for administering a composition containing one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be any duration that effectively treats a cancer present within the mammal without producing significant toxicity to the mammal. In some cases, the effective duration can vary from several months to several years. In general, the effective duration for treating a mammal having a cancer can range in duration from about one or two months to five or more years. Multiple factors can influence the actual effective duration used for a particular treatment.

For example, an effective duration can vary with the frequency of administration, effective amount, use of multiple treatment agents, route of administration, and severity of the condition being treated.

In certain instances, a cancer within a mammal can be monitored to evaluate the effectiveness of the cancer treatment. Any appropriate method can be used to determine whether or not a mammal having cancer is treated. For example, imaging techniques or laboratory assays can be used to assess the number of cancer cells and/or the size of a tumor present within a mammal. For example, imaging techniques or laboratory assays can be used to assess the location of cancer cells and/or a tumor present within a mammal.

In some cases, one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4) can be administered to a mammal having a cancer as a combination therapy with one or more additional cancer treatments ( e.g ., anti-cancer agents). A cancer treatment can include any appropriate cancer treatments. In some cases, a cancer treatment can include surgery. In some cases, a cancer treatment can include radiation therapy. In some cases, a cancer treatment can include administration of one or more therapeutic agents (e.g., one or more anti-cancer agents). Examples of anti-cancer agents include, without limitation, platinum compounds (e.g., a cisplatin or carboplatin), taxanes (e.g, paclitaxel, docetaxel, or an albumin bound paclitaxel such as nab-paclitaxel), altretamine, capecitabine, cyclophosphamide, etoposide (vp-16), gemcitabine, ifosfamide, irinotecan (cpt-11), liposomal doxorubicin, melphalan, pemetrexed, topotecan, vinorelbine, luteinizing-hormone-releasing hormone (LHRH) agonists (e.g, goserelin and leuprolide), anti-estrogens (e.g, tamoxifen), aromatase inhibitors (e.g, letrozole, anastrozole, and exemestane), angiogenesis inhibitors (e.g, bevacizumab), poly(ADP)-ribose polymerase (PARP) inhibitors (e.g, olaparib, rucaparib, and niraparib), radioactive phosphorus, anti- CTLA-4 antibodies, anti -PD- 1 antibodies, anti-PD-Ll antibodies, IL-2 and other cytokines, other bispecific antibodies, and any combinations thereof. In cases where one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein are used in combination with one or more additional cancer treatments, the one or more additional cancer treatments can be administered at the same time or independently. For example, a composition including one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein can be administered first, and the one or more additional cancer treatments administered second, or vice versa.

Also provided herein are kits that include one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein (e.g, a modified peptide including an amino acid sequence set forth in any one of SEQ ID NO: 1, SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4). For example, a kit can include a composition (e.g, a pharmaceutically acceptable composition) containing one or more molecules including one or more antigen-binding domains (e.g, scFvs) that can bind to a modified peptide described herein. In some cases, a kit can include instructions for performing any of the methods described herein. In some cases, a kit can include at least one dose of any of the compositions (e.g, pharmaceutical compositions) described herein. In some cases, a kit can provide a means ( e.g ., a syringe) for administering any of the compositions (e.g., pharmaceutical compositions) described herein.

The invention will be further described in the following examples, which do not limit the scope of the invention described in the claims. EXAMPLES

Example 1: Identification ofMANAbody Clones and Conversion ofMANAbody Clones into T cell-based Therapeutic Formats

In this study, a phage display library was designed and built, which displayed a single chain variable fragment (scFv) on the phage surface. The scFvs present in the library were based on the humanized 4D5 (trastuzumab) framework with amino acid variability introduced at key positions of the scFv’s complementarity determining regions (CDRs).

Phage display library was used to identify scFvs that specifically recognized mutation-containing peptides folded into a complex with a recombinant HLA allele alpha chain and beta-2 microglobulin (b2M). These complexes, also referred to herein as monomers, mimic the natural peptide/HLA complexes on a cancer cell surface.

Peptide-HLA targets can include mutant peptides (e.g, MANAs) shown in Table 1. scFvs that can specifically bind to peptide-HLA targets in Table 1 are shown in Table 3. These scFvs can also be referred to as MANAbodies for their ability to bind to MANAs.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 3. MANAbody scFv sequences.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Flow cytometry data for scFvs that specifically recognized a p53 peptide containing the R175H mutation in complex with HLA-A2 (HMTEVVRHC; SEQ ID NO: 1) are shown in Figure 1. The scFvs specifically stained the HLA allele-matched cell lines when these cells are pulsed with the mutant peptide, but not the WT peptide or not pulsed with peptide at all. Flow cytometry data for scFvs specific for H/K/N RAS Q61H, Q61L, and Q61R peptides are shown in Figures 2-4.

To demonstrate that MANAbody clones can be utilized as a therapeutic modality, selected MANAbody clones were engineered into bispecific antibodies having one antibody- fragment binding to a mutant peptide presented in the context of HLA and having one antibody-fragment binding to a CD3 protein on the T cell surface (Table 4). The bispecific antibodies are referred to as single-chain diabody (scDb) hereafter. Specifically, bispecific antibodies targeting a mutant p53 R175H peptide presented in the context of HLA-A2 and CD3 were engineered, and bispecific antibodies targeting mutant H/K/N RAS Q61H, Q61L, and Q61R peptides presented in the context of HLA-Al and CD3 were engineered.

Representative scDb co-culture results are shown in Figure 5 for five p53 R175H HLA-A2 MANAbody scFv clones combined with anti-CD3 scFv. T cells were co-cultured with COS-7 cells co-transfected with plasmids encoding HLA-A2, full-length p53 variants, and/or GFP in the presence of the specified concentration of scDb. As a read out of T cell activation by cognate antigen on target cells, the release of IFNy in the co-culture media supernatant was measured by ELISA. Only when COS-7 cells were co-transfected with HLA-A2 and mutant p53 R175H plasmids was there significant T cell release of IFNy over background, with the level of IFNy dependent on the concentration of scDb included in the well. T cells co-cultured with COS-7 cells co-transfected with HLA-A2 and WT p53 released only background levels of IFNy. Representative scDb co-culture results are shown in Figures 6-8 for H/K/N RAS Q61H, Q61L, and Q61R HLA-Al MANAbody scFv clones combined with an anti-CD3 clone into a scDb. In these co-cultures, only when COS-7 cells were co-transfected with HLA-Al and mutant full-length KRAS Q61H, Q61L, or Q61R plasmids was there significant T cell release of IFNy over background. T cells co-cultured with COS-7 cells co-transfected with HLA-Al and WT KRAS released only background levels of IFNy, similar to the levels of IFNy seen in no scDb wells. To evaluate whether the H/K/N RAS Q61H, Q61L, and Q61R scDbs can react against all 3 RAS isoforms with the cognate Q61 mutations, COS-7 cells were transfected with HLA-A1 and plasmids encoding for full-length HRAS, KRAS, or NRAS that are WT, or harbored Q61H, Q61K, Q61L, or Q61R mutation. The transfected COS-7 cells were then co-cultured with T cells and representative H/K/N RAS Q61H, Q61L, and Q61R scDbs. Figure 9a shows that the H/K/N RAS Q61H scDb only induced IFNy in the presence of COS-7 cells co-transfected with HLA-Al and the Q61H mutant HRAS, KRAS, or NRAS. Likewise, H/K/N RAS Q61L scDb only induced PTNg in the presence of COS-7 cells co transfected with HLA-Al and the Q61L mutant HRAS, KRAS, or NRAS (Figure 9b) and H/K/N RAS Q61R scDb only induced IFNy in the presence of COS-7 cells co-transfected with HLA-Al and the Q61R mutant HRAS, KRAS, or NRAS (Figure 9c).

To evaluate the ability of MANAbody clones to recognize tumor cells, tumor cell lines with endogenous cognate HLA and mutations were co-cultured with T cells and scDbs. An endogenous p53 R175H HLA-A2 positive cell line TYKnu, along with its isogenic p53 knockout control, was cultured with T cells and p53 R175H HLA-A2 scDb. IFNy release was only induced against the parental TYKnu cell line but not the p53 knockout TYKnu (Figure 10). An endogenous NRAS Q61L HLA-Al positive cell line HL-60, along with its isogenic HLA-Al knockout control, was cultured with T cells and H/K/N RAS Q61L HLA- Al scDb. IFNy release was only seen against the parental HL-60 cell line but not the HLA- Al knockout HL-60 (Figure 11). Together, these findings suggest that bispecific antibodies containing MANAbody clones can target tumor cells expressing MANAs presented in the context of HLA molecules.

To evaluate the efficacy of using MANAbody clones as a therapeutic modality, target cell viability of p53 R175H HLA-A2 and H/K/N RAS Q61L HLA-Al scDb co-cultures was assayed using Promega’s CellTiter-Glo reagent. CellTiter-Glo measures ATP concentration in a well, which is proportional to the number of viable cells. Percent target cell viability was measured by subtracting the CellTiter-Glo value from T cell only wells and normalizing to target cell only wells. Only when parental TYKnu cells were incubated with T cells in the presence of the p53 R175H HLA-A2 scDb, was there significant target cell death (Figure 12). No target cell death was observed among the TYKnu p53 knockout wells. Similarly, only when parental HL-60 cells were incubated with T cells in the presence of the H/K/N RAS Q61L HLA-A1 scDb, was there significant target cell death (Figure 13). No target cell death was observed among the HL-60 HLA-A1 knockout wells.

Together, these findings demonstrate that MANAbodies can be used to redirect and activate T cells to kill tumor cells expressing particular mutant protein and HLA allele pairs (e.g., p53 R175H with HLA-A2 and H/K/N RAS Q61L with HLA-A1).

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 4. Anti-human CD3 scFv sequences

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 5. Anti-human CD16a scFv sequences

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Materials and Methods

Cells and Cell Lines.

RPMI-6666 cells (ATCC, Manassas, VA) were cultured in RPMI-1640 (ATCC) with 20% FBS (GE Hyclone, Logan, Utah, USA), and 1% penicillin streptomycin (Life Technologies). T2 cells (ATCC) and TYKnu (JCRB, Japan) were cultured in RPMI-1640 (ATCC) with 10% FBS (GE Hyclone), and 1% penicillin streptomycin (Thermo Fisher). SigM5 cells (DSMZ, Brunswick, Germany) and HL-60 cells (ATCC) were cultured in Iscove's MDM (ATCC) with 20% FBS (GE Hyclone), and 1% penicillin streptomycin (Thermo Fisher). COS-7 cells (ATCC) was cultured in McCoy's 5A (Modified) Medium (Thermo Fisher) with 10% FBS (GE Hyclone), and 1% penicillin streptomycin (Thermo Fisher). COS-7 cells (ATCC, CRL-1651™) were cultured in DMEM (high glucose, pyruvate; Thermo Fisher) with 10% FBS (GE Hyclone), and 1% Penicillin-Streptomycin (Thermo Fisher). 293FT cells (Thermo Fisher) were cultured in high-glucose D-MEM (Thermo Fisher), with 10% FBS (GE Hyclone), 0.1 mM MEM Non-Essential Amino Acids (NEAA, Thermo Fisher), 6 mM L-glutamine (Thermo Fisher), 1 mM MEM Sodium Pyruvate (Thermo Fisher), 500 pg/ml geneticin (Thermo Fisher), and 1% Penicillin- Streptomycin (Thermo Fisher). All cell lines were maintained at 37°C under 5% CO2.

PBMCs were obtained by Ficoll-Paque PLUS (GE Healthcare) gradient centrifugation of whole blood from healthy volunteer donors. To activate and expand T cells, PBMCs were cultured with 15 ng/ml OKT3 (BioLegend, San Diego, CA), 100 IU/mL recombinant human interleukin-2 (Aldesleukin, Prometheus Therapeutics and Diagnostics, San Diego, CA), and 5 ng/ml recombinant human interleukin-7 (BioLegend, San Diego, CA) in RPMI-1640 (ATCC) with 10% FBS (GE Hyclone), 1% Penicillin-Streptomycin (Life Technologies) at 37°C under 5% CCkfor 3 days. After 3 days, the expanded T cells were kept in the same cytokine-containing media without OKT3.

Phage Display Library Construction.

Oligonucleotides were synthesized by GeneArt (Thermo Fisher Scientific) using trinucleotide mutagenesis (TRIM) technology. The oligonucleotides were incorporated into the pADL-lOb phagemid (Antibody Design Labs, San Diego, CA). This phagemid contains an FI origin, a transcriptional repressor to limit uninduced expression, a lac operator, and a lac repressor. The scFv was synthesized with a pelB periplasmic secretion signal and was subcloned downstream of the lac operator. A FLAG (DYKDDDDK; SEQ ID NO: 190) epitope tag was placed immediately downstream of the variable heavy chain, which was followed in frame by the full-length M13 pill coat protein sequence.

Ten ng of the ligation product was mixed on ice with 10 pL of electrocompetent SS320 cells (Lucigen, Middleton, WI) and 14 pL of double-distilled water. This mixture was electroporated (200 ohms, 25 microFarads, 1.8kV) using a Gene Pulser electroporation system (Bio-Rad, Hercules, CA) and allowed to recover in Recovery Media (Lucigen) for 45 minutes at 37°C. Cells transformed with 60 ng of ligation product were pooled and plated on a 24-cm x 24-cm plate containing 2xYT medium supplemented with carbenicillin (100 pg/mL) and 2% glucose. Cells were grown at 37°C for 6 hours and placed at 4°C overnight. To determine the transformation efficiency for each series of electroporations, aliquots were taken and titered by serial dilution. Cells grown on plates were scraped into 850 mL of 2xYT medium with carbenicillin (100 pg/mL) plus 2% glucose for a final Oϋόoo of 5-15. Two mL of the 850 mL culture were taken and diluted ~1 :200 to reach a final Oϋόoo of 0.05-0.07. To the remaining culture, 150 mL of sterile glycerol were added before snap freezing to produce glycerol stocks. The diluted bacteria were grown to an Oϋόoo of 0.3-0.5, infected with M13K07 Helper phage at an MOI of 4 (Antibody Design Labs) and shaken at 37°C for 1 hour. The culture was centrifuged and the cells were resuspended in 2xYT medium with carbenicllin (100 pg/mL), kanamycin (50 pg/mL), and IPTG (50 mM, Thermo Fisher) and grown overnight at 30°C for phage production. The following morning, the bacterial culture was aliquoted into 50 mL Falcon tubes and pelleted twice at high speed to obtain clarified supernatant. The phage-laden supernatant was precipitated on ice for 40 minutes with a 20% PEG-8000/2.5M NaCl solution at a 4:1 ratio of PEG/NaCl:supernatent. After precipitation, phage was centrifuged at 12,000 g for 40 minutes and resuspended in lx TBS, 2 mM EDTA, and lx Complete Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO). Phages from multiple tubes were pooled and re-precipitated. The total number of transformants obtained was determined to be 3.6 x 10¹⁰. The library was aliquoted and stored in 15% glycerol at - 80°C. Next-generation sequencing of the complete phage library.

DNA from the libraries was amplified using primers that flank the CDR-H3 region. The sequences at the 5'-ends of these primers incorporated molecular barcodes to facilitate unambiguous enumeration of distinct phage sequences. The protocols for PCR-amplification and sequencing are described by Kinde et al. (2011 PNAS. 108:9530-35). Sequences processed and translated using a custom SQL database and both the nucleotide sequences and amino acid translations were analyzed using Microsoft Excel.

Peptides and HLA-Monomers.

Mutant and WT peptides (listed in Table 1) were predicted to bind to HLA alleles using NetMHC version 4.0. All peptides were synthesized at a purity of >90% by Peptide 2.0 (Chantilly, VA). Peptides were resuspended in DMSO or DMF at 10 mg/mL and stored at -20°C. HLA monomers were synthesized by refolding recombinant HLA with peptide and beta-2 microglobulin, purified by gel-filtration, and biotinylated (Fred Hutchinson Immune Monitoring Lab, Seattle, WA). Monomers were confirmed to be folded prior to selection by performing an ELISA using W6/32 antibody (BioLegend, San Diego, CA).

Selection for phage binding to mutant peptide-HLA monomers. scFv-bearing phage clones specific to the mutant pHLA complexes were identified similar to methods described elsewhere (see, e.g., Skora etal, 2015 PNAS.112:9967-72). The panning schema involved an enrichment phase, a competition phase, and a final selection phase.

The phage display library stored at -20°C in 15% glycerol, were regrown within a week of starting the panning process. A colony of phage-competent SS320 cells (Lucigen, Middleton, WI) was inoculated in a 37°C overnight culture of 2xYT medium (Sigma- Aldrich, St. Louis, MO) supplemented with tetracycline (20 pg/mL), and the next day grown to 2 L of mid-log phase (Oϋόoo of 0.3-0.5) bacteria. Bacteria were infected with the phage library at an MOI of 0.5 and M13K07 Helper phage (Antibody Design Labs, San Diego, CA) at an MOI of 4 along with the addition of 2% (W/V) glucose (Sigma-Aldrich, St. Louis, MO) and allowed to shake for 1 hour at 37°C. The culture was centrifuged and the cells were resuspended in 2xYT medium with carbenicillin (100 pg/mL), kanamycin (50 pg/mL), and 50 pM IPTG and subsequently shaken and grown overnight at 30°C for phage production. The following morning, the bacterial culture was aliquoted into 50 mL Falcon tubes and pelleted twice at high speed to obtain clarified supernatant. The phage-laden supernatant was precipitated on ice for 40 minutes with a 20% PEG-8000/2.5M NaCl solution at a 1 :4 ratio of PEG/NaCl: supernatant. After precipitation, phage was centrifuged at 12,000 x g for 40 minutes and resuspended in 1 mL of IX TBS with 2 mM EDTA, 0.1% sodium azide, and lx Complete Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO).

Biotinylated pHLA monomer complexes were conjugated to M-280 streptavidin magnetic Dynabeads (Life Technologies, Carlsbad, CA). The biotinylated pHLA were incubated with either 25 pL of Dynabeads beads per 1 pg of pHLA in blocking buffer (PBS, 0.5% BSA, 2mM EDTA, and 0.1% sodium azide) for 1 hour at room temperature (RT).

After the initial incubation, the complexes were washed and resuspended in 100 pL of blocking buffer.

During the enrichment phase (Round 1), approximately 2 x 10¹³ phage, representing ~500-fold coverage of the library, were negatively selected for 1 hour at RT with a mixture of 1 mL unconjugated washed Dynabeads, 1 mg free streptavidin protein (RayBiotech, Norcross, GA) to remove any phage recognizing either the unconjugated Dynabead and streptavidin. After negative selection, beads were isolated with a DynaMag-2 magnet (Life Technologies, Carlsbad, CA) and the supernatant containing unbound phage was transferred for positive selection for 1 hour at RT against the 1 pg of mutant pHLA conjugated to Dynabeads. Prior to elution, beads were washed 10 times with 1 mL of TBST (lx TBS with 0.5% Tween-20). Phage were eluted by resuspending the beads in 1 mL of 0.2 M glycine, pH 2.2. After a 10 minutes incubation, the solution was neutralized by the addition of 150 pL of 1 M Tris, pH 9.0. Neutralized phages were used to infect 10 mL cultures of mid-log-phage SS320s, with the addition of M13K07 helper phage (MOI of 4) and 2% glucose. Bacteria were then incubated as previously described and the phages were precipitated the next morning with PEG/NaCl.

During the selection phase (Rounds 2-5), phage from the previous round was subjected to negative selection against HLA-matched cell lines without the mutations of interest, corresponding WT pHLA monomer conjugated to Dynabeads, unrelated pHLA monomerconjugated to Danybeads, and free streptavidin. After negative selection, beads were isolated with a DynaMag-2 magnet and unbound phage was transferred for positive selection. This was performed by incubating phage with 1 pg (Round 2), 0.5 pg (Rounds 3, 4), or 0.25 pg (Round 5) mutant pHLA conjugated to the magnetic Dynabeads. Prior to elution, beads were washed 10 times in 1 mL TBST. Phage were eluted from magnetic Dynabeads and used to infect mid-log phase SS320 cells as described above.

Flow Cytometry.

Monoclonal phage flow cytometry staining was performed by selecting individual colonies of SS320 cells transformed with a limiting dilution of phage obtained from the final selection round. Individual colonies were inoculated into 200 pi of 2xYT medium containing 100 pg/mL carbenicillin and 2% glucose and grown for three hours at 37°C. The cells were then infected with 1.6 x 10⁷ M13K07 helper phage and incubated for 1 hour at 37°C with shaking. The cells were pelleted, resuspended in 300 pL of 2xYT medium containing carbenicillin (100 pg/mL), kanamycin (50 pg/mL), and 50 pM IPTG, and grown overnight at 30°C for phage production. Cells were pelleted and the phage-laden supernatant was used for staining.

For peptide pulsing, HLA-matched cells were washed once with PBS and once with serum-free RPMI-1640 before incubation at 10⁶ cells per mL in serum-free RPMI-1640 containing 50 pg/mL peptide and 10 pg/mL human beta-2 microglobulin (ProSpec, East Brunswick, NJ) overnight at 37°C. The pulsed cells were pelleted, washed once in cold staining buffer (PBS containing 0.5% BSA, 2 mM EDTA, and 0.1% sodium azide), and resuspended in 50 pL of stain buffer. Phage staining was performed on ice with 50 pL monoclonal phage supernatant for 1 hour, followed by one 800 pL wash in cold staining buffer. Cells were then stained with lpg of rabbit anti-M13 antibody (Novus Biologies, Centennial, CO) in 100 pL total volume on ice for 1 hour and washed once with 800 pL of cold staining buffer. Cells were then stained with 5 pL anti-rabbit-PE (Biolegend) on ice for 1 hour in 100 pL total volume, followed by incubation with 200 pL LIVE/DEAD Fixable Near-IR Dead Cell Stain (Thermo Fisher) for 10 minutes at room temperature per manufacturer’s instructions. Cells were washed once in 800 pL of staining buffer. Stained cells were analyzed using an iQue Screener (IntelliCyt, Albuquerque, NM). Bispecific Antibody Production. gBLOCKs encoding bispecific antibodies were ordered from IDT (Skokie, Illinois). gBLOCKs were cloned into the pcDNA3.4 plasmid (Thermo Fisher) by NEBuilder HiFi DNA Assembly (New England Biolabs, Ipswich, MA) following the manufacturer’s protocol. 293FT cells (Thermo Fisher) were transfected with the bispecific antibody pcDNA3.4 plasmids using Lipofectamine 3000 (Life Technologies) per the manufacturer’s instructions in a T75 flask. Following a 5-8 day incubation, media was harvested and centrifuged at 500g for 10 minutes at 4°C. Bispecific antibody protein was purified using a Clontech Capturem™ His-Tagged Purification Mixiprep Kit (Takara Bio, Mountain View, CA) per manufacturer’s instructions. Bispecific antibody protein was desalted into PBS using Zeba spin 7k MWCO desalting columns per the manufacturer’s instructions.

Bispecific antibody concentration was quantified using Mini-PROTEAN® TGX Stain- Free™ Precast Gels (Biorad, Hercules, California) using a standard curve of protein of known concentration. Stain-free gels were imaged using the ChemiDoc XRS+ Imager (Biorad).

Bispecific Antibody Co-Culture Assay.

COS-7 cells were transfected with various combinations of pcDNA3.1 or pcDNA3.4 (Life Technologies) plasmids encoding HLA-A2, HLA-A1, p53(WT, R175H), HRAS(WT, Q61H, Q61K, Q61L, Q61R), KRAS(WT, Q61H, Q61K, Q61L, Q61R), NRAS(WT, Q61H, Q61K, Q61L, Q61R) with Lipofectamine 3000 (Life Technologies) per manufacturer’s instructions in a T75 flask. A total of 50,000 T cells were combined with transfected 50,000 COS-7 cells, 25,000 TYKnu cells, or 25,000 HL-60 cells and the specified concentration of bispecific antibody in a 96-well plate, and the co-culture was allowed to incubate for 20 hours at 37°C under 5% CO2. Following co-culture, conditioned media was collected and assayed for secreted IFNy by Quantikine® ELISA (R&D Systems, Minneapolis, MN). Alternatively, following coculture, target cell viability was measured using CellTiter-Glo (Promega, Madison, WI) per the manufacturer’s instructions.

CRISPR Cell Line Engineering.

The Alt-R CRISPR system (Integrated DNA Technologies, IDT) was used to modify the p53 of the TYKnu cell line and the HLA allele of HL-60 cell line. Alt-R® CRISPR Cas9 crRNAs (IDT) targeting TP 53 exon 3 (CCCCGGACGATATTGAACAA; SEQ ID NO: 191), and HLA-A exon 2 (CAGACTGACCGAGCGAACCT; SEQ ID NO: 192) as well as Alt-R® CRISPR-Cas9 tracrRNA (IDT) were resuspended at 100 mM with Nuclease-Free Duplex Buffer (IDT). The crRNAs and tracrRNA were duplexed at a 1:1 molar ratio for 5 minutes at 95°C according to the manufacturer’s instructions. The duplexed RNA was allowed to cool to room temperature prior to mixing with Cas9 Nuclease (IDT) at a 1.2: 1 molar ratio for 15 minutes. To knock out p53 of TYKnu cells, 40 pmols of the Cas9 RNP complexed with p53 gRNA were mixed with 200,000 TYKnu cells in 20 pL of OptiMEM. This mixture was loaded into a 0.1 cm cuvette (Biorad) and electroporated at 120V and 16ms using an ECM 2001 (BTX). To knock out the HLA-A alleles in HL-60 cells, 40 pmols of the Cas9 RNP complexed with HLA-1 gRNA were mixed with 200,000 HL-60 cells in 20 pL of OptiMEM. This mixture was loaded into a 0.1 cm cuvette (Biorad) and electroporated at 150V and 16 ms using an ECM 2001 (BTX). After electroporation, cells were immediately transferred to complete growth medium and cultured for 10 days, changing media and passaging as needed. p53 and HLA-A modified polyclonal pools were plated at a density of 0.5 cells per well in 96 well plates and cultured for 2 weeks. Single colonies were harvested and plated into 2 replicate 96-well plates. Genomic DNA was harvested from one of the plates using the Quick-DNA™ 96 Kit (Zymo Research, Irvine, CA), PCR amplified using Q5® Hot Start High-Fidleity 2X Master Mix (New England Biolabs), and Sanger sequenced (Genewiz, South Plainfield, NJ) to select for clones with the desired modifications.

Example 2: Targeting a Neoantigen Derived From a common TP53 Mutation

TP 53 is the most commonly mutated cancer driver gene, but despite extensive efforts, no drug targeting mutant TP53 has been approved for treatment of the large number of patients whose tumor contain p53 mutations. This Example describes the identification of an antibody highly specific to the most common TP53 mutation (R175H) in complex with a common HLA-A allele on the cell surface. For example, this Example describes the identification of a TCRm antibody specific to the HLA-A*02:01-restricted p53^R175H neoantigen, the structural basis of its specificity, and its conversion to a bispecific antibody that can lyse cancer cells in a fashion dependent on the presence of the neoantigen. Such an immunotherapeutic agent that targets a common TP53 mutation can be used to target cancers containing other tumor suppressor gene mutations.

Results

The p53^R175H neoantigen is presented on the surface of cancer cells The p53^R175H (aa 168-176, HMTEVVRHC; SEQ ID NO:l) and p53^WT

(HMTEVVRRC; SEQ ID NO: 135) peptides were predicted on the NetMHCpan 4.0 server to bind HLA-A*02:01 at 5177.6 nM (rank 9.6%) and 7121.5 nM (11.6%), respectively. To provide experimental evidence of and to quantify such presentation, peptides eluted from HLA molecules were analyzed in four different cell culture systems using a mass spectrometry (MS)-based method. First, the human HLA-A*02:01 and either p53^R175H or p53^WT were co-expressed in monkey COS-7 cells. MS analysis of the peptides immunopurified with an anti-HLA antibody detected the p53^R175H peptide at approximately 1500 copies per cell (Fig. 20, Table 6). Though relatively abundant p53^R175H peptide was detected, the p53^WT peptide was not observed. Second, MS analysis was performed on three human cancer cell lines, KMS26, TYK-nu, and KLE, all of which harbor p53^R175H mutations and carry an HLA-A*02:01 allele. The p53^R175H peptide was detected on all three cell lines, and, as expected, at much lower levels than in the COS-7 cells in which the mutant TP53 and HLA genes were exogenously introduced (Fig. 20, Table 6). Based on comparisons with heavy isotope labeled controls, it was estimated that there were 2.4, 1.3, and 1.5 copies of cell-surface p53^R175H/HLA-A*02:01 complexes on the cell surfaces of KMS26, TYK-nu, and KLE cell lines, respectively (Table 6).

Table 6. Quantitative assessment of the p53^R175H neoantigen peptide. The amount of p53^R175H neoantigen peptide (HMTEVVRHC; SEQ ID NO:l) present in COS-7 cells transfected with HLA-A*02:01 and p53^R175H or p53^WT, as well as cells lines that endogenously express HLA-A*02:01 and p53^R175H, were quantified using mass spectrometry.

Identification of scFv-expressing phage clones specific for the HLA-A*02: 01 -restricted p53^R175H peptide and conversion to scDb format

To identify TCR-mimic single-chain variable fragments (scFvs) selectively targeting mutant pHLA complexes, an scFv-displaying phage library was screened with an estimated complexity >lxl0¹⁰. Positive selection against HLA-A*02:01 pHLA monomers containing the p53^R175H peptide were combined with negative selection against pHLA monomers containing the p53^WT and irrelevant peptides. Selected phage clones were amplified and assessed for binding to T2 cells presenting the mutant or wild-type (WT) peptide via flow cytometry (Fig. 21).

Twenty-three phage clones with median fluorescence intensity (MFI) ratios of p53^R175H to p53^WT >4 were then converted to T cell-retargeting bispecific antibodies (Fig. 21). This was achieved through linking each individual scFv to an anti-CD3 scFv (UCHT1) in a single-chain diabody (scDb) format (Fig. 22). The scDb format was chosen after evaluating several previously described bispecific antibody formats, such as bispecific T-cell engager (BiTE), dual-affinity re-targeting antibody (DART), and diabody in pilot experiments. The ability of scDbs to activate T cells was assessed by interferon-g (IFN-g) release after co-incubation with COS-7 cells overexpressing HLA-A*02:01 and either full- length p53^R175H or p53^WT proteins. Two scDb clones, named H2-scDb and H20-scDb and derived from phage clones H2 and H20, respectively, showed the most potent and specific T- cell activation in the presence of p53^R175H/HLA-A*02:01 (Fig. 23). The specificity of these scDbs was further evaluated by titration enzyme-linked immunosorbent assay (ELISA).

Both H2 and H20 bound to p53^R175H/HLA-A*02:01 at low concentrations, as expected. At high concentrations, H20-scDb also bound to p53^WT/HLA-A*02:01, while H2-scDb did not bind to the wild type pHLA complex even at very high concentrations of the scDb (Fig. 1 A, Fig. 24). H2-scDb was therefore chosen for further analysis. As assessed by surface plasmon resonance (SPR), the H2-scDb bound to p53^R175H/HLA-A*02:01 with a KD=%6 nM, a kon of 1.76 x lCPM ¹ S^'1, and a k₀ff of 1.48 x 10² s ¹ (Fig. 14B). The k₀n of 1.76 x 10⁵ NT ¹ s ¹ suggested a lack of conformational change of the p53^R175H/HLA-A*02:01 upon binding. No detectable binding of the H2-scDb to p53^WT/HLA-A*02:01 was observed in the SPR experiments (Fig. 14B).

Next, it was examined whether anti-CD3 arms of the scDb other than the original UCHT1, could influence the ability of H2 to induce T-cell activation. The H2 scFv was linked to a panel of commonly used anti-CD3r scFvs, including UCHT1, hUCHTlv9, OKT3, TR66, and hXR32. It was found that, among the anti-CD3 scFvs tested, UCHT1 activated T cells at the lowest p53^R175H peptide concentration when linked to the H2 scFv (Fig. 25), and used this particular bispecific antibody for further experiments.

H2-scDb specifically recognizes cancer cells expressing the p53^R175H neoantigen

It was next evaluated the ability of H2-scDb to recognize cancer cell lines harboring the p53^R175H mutations and expressing various levels of HLA-A* 02:01. H2-scDb elicited T- cell responses in a dose-dependent manner when T cells were co-cultured with three lines that expressed moderate to high levels of HLA-A*02:01 (Fig. 15A). This activation was noted even at very low (sub-nanomolar) concentrations of the bispecific antibody. The T-cell responses were polyfunctional, as indicated by the release of cytotoxic granule proteins granzyme B and perforin, cytotoxicity, and the production of cytokines IFN-g, tumor necrosis factor a (TNF-a), interleukin-2 (IL-2), and others (Fig. 15B, Fig. 26). Clustering of T cells around tumor cells, leading to their lysis in the presence of H2-scDb, was also visualized by real-time live-cell imaging (Fig. 15C). The specificity of the bispecific antibody for both the p53^R175H peptide and HLA-A*02:01 was evident from the observation that much lower levels of IFN-g were induced by cells harboring a p53^R175H mutation but low levels of expression of HLA-A*02:01 (AU565 or SK-BR3) or by cells without p53^R175H but relatively high levels of HLA-A*02:01expression (Fig. 15A, Fig. 27).

It was further validated the specificity of H2-scDb using nine pairs of isogenic cell lines that differed with respect to HLA-A*02:01 expression or p53^R175H mutation. First, human HEK293FT (77^J53^,r//HLA-A*02:0 l ) or Saos-2 (77^J53"^!,///HLA-A*02:0 l ) cells were transfected with plasmids expressing either full-length p53^R175H or p53^WT. H2-scDb induced robust T-cell activation when co-cultured with both cell lines overexpressing p53^R175Hbut not with p53^WT-overexpressing or parental cells (Fig. 16A). Second, HLA-A*02:01 -encoding retrovirus were transduced into four cell lines (AU565, SK-BR-3, HuCCTl, CCRF-CEM) that harbored the p53^R175H mutation but had low levels of HLA-A*02:01 expression (Fig.

28). Exogenous expression of HLA-A*02:01 in all four lines conferred T-cell activation by H2-scDb (Fig. 16B). Third, TP53 was genetically disrupted in KMS26, TYK-nu, and KLE cancer cell lines that carry endogenous p53^R175H using a CRISPR-based technology (Fig. 29A). T-cell activation, as assessed by IFN-g secretion, was reduced to control levels when TP53 was knocked out in all three cell lines (Fig. 16C). The cytotoxicity mediated by H2- scDb was similarly mitigated by TP 53 knock-out (KO) in these cells (Fig. 16D, Fig. 29B).

Overall structure of the H2-Fab p53^R175n /HLA-A *02:01 ternary complex

To understand the structural basis for the high specificity of the H2 clone for p53^R175H/HLA-A*02:01, the H2 fragment antigen-binding (H2-Fab)-p53^R175H/HLA-A*02:01 complex was purified (Fig. 30) and its crystal structure was determined by molecular replacement and refined to 3.5 A resolution (PDB ID 6W51, Table 7). There were four H2- Fab and four p53^R175H/HLA-A*02:01 per asymmetric unit (Fig. 17A, B). All four H2-Fab were firmly positioned on the p53^R175H/HLA-A*02:01, without evidence of rocking with a root-mean-square deviation (rmsd) of 0.45 to 0.51 A. The total buried surface area of the H2-Fab-p53^R175H/HLA-A*02:01 interface was 1173 A², with roughly equal contributions from the heavy and light chains (644 A² and 529 A², respectively, Table 8). Although the entire structure was refined to a resolution of 3.5 A, particularly clear electron densities were observed for the p53^R175H neoantigen, the CDRs of the H2-Fab, and the HLA-A*02:01 (Fig. 17C, D). Table 7. X-ray Crystallography data collection and refinement statistics.

Binding of the p53^R175H peptide to HLA-A *02:01

The p53^R175H neoantigen occupies the binding cleft al-a2 of HLA-A* 02:01, burying a solvent accessible surface area of 870 A², slightly larger than other peptide/HLA-A*02:01 complexes (Fig. 18A, B, Fig. 31 A) and with the C-terminal arginine at position 7 (P7, Arg7, Argl74 in p53) and histidine at position 8 (P8, His8, Hisl75 in mutant p53) pointing up, out of the groove. In contrast, the N-terminus of the peptide is situated deep within the peptide binding cleft, anchored by multiple residues in the HLA-A*02:01 (Fig. 18A, B, Fig. 31 A). The anchor residues of the peptide, a methionine at position 2 (P2, Met2, Metl69 in p53) and a cysteine residue at position 9 (P9) (Fig. 3 IB), departed from the canonical anchor residues — leucine at P2 and valine or leucine at P9. Peptides that bind to HLA-A*02:01 through either a methionine at P2 or a cysteine at P9 have been reported, but not both. Based on alignments with structures of other HLA-A*02:01 peptides in complex with TCR or TCRm, the unconventional anchoring of p53^R175H did not result in drastic peptide conformational change or positioning (Fig. 31C, D).

Structural basis for the recognition of p53^R17m/HLA-A*02:01 by the H2-Fab

The recognition of the HLA-A*02:01 by the H2-Fab was mediated by all six CDRs. There were a total of 79 contacts between the H2-Fab CDRs and the al and a2 of HLA- A*02:01, with the light chain contributing to 61% of those contacts. The H2-Fab buried a solvent accessible surface area of 818 A² within the HLA, of which 427 A² were contributed by the light chain and 391 A² by the heavy chain (Table 8). In contrast, only four of the six H2-Fab CDRs (HI, H2, H3 and L3) interacted with the p53^R175H peptide. Overall, the H2- Fab made 36 contacts with the p53^R175H neoantigen, including five hydrogen bonds and numerous van der Waals interactions. Importantly, Hisl75 at P8 made 47% of all direct contacts with the H2-Fab. The CDR-H1, H2, and H3 of the heavy chain and CDR-L3 of the light chain formed a cage-like configuration around the C-terminus of the p53^R175H peptide, trapping Argl74 at P7 and Hisl75 at P8 into position by providing a stable interaction (Fig. 18C). The imidazole side chain of Hisl75 at P8 was anchored by a hydrogen bonding network with Asp54 (CDR-H2) and Tyr94 (CDR-L3) (Fig. 18C, Fig. 32). Tyr52 (CDR-H2) acted as a ceiling and capped the cage-like structure around His8 by forming p-p interactions (Fig. 18C, Fig. 32).

Table 8. Structural comparison of H2-Fab-p53^R175H/ HLA-A*02:01 with various TCR and Fab antibody-pHLA. Total bonds were calculated using a 4 A cutoff which includes both hydrogen bonds and van der Waals. PDB, Protein Data Bank; BSA, buried surface area; a, TCRa chain; b, TCRP chain; H, VH domain; L, VL domain; pep, HLA presented peptide.

Viewed from the axis of the C-terminus to the N-terminus of the p53^R175H peptide, the CDRs were arranged in the order H2, HI, L3, H3, LI, L2 (Fig. 17E, F, G). Interestingly, the axis of the H2-Fab was nearly parallel to the axis of the peptide within the HLA groove with a binding angle of 27° (Fig. 17G, H). This orientation was quite different from that of most previously described TCRs or TCRm antibodies to pHLA complexes, in which the axes are diagonal (Fig. 33).

Assessing candidate cross-reactive peptides

One of the major challenges confronting new immunotherapeutic antibodies is off- target binding, which can result in toxicity to normal cells. Several powerful approaches to profile TCR and TCRm specificity have been developed to address this important issue. Scanning mutagenesis was employed to identify peptides in the human proteome to which H2-scDb might cross-react. A peptide library was generated by systemically substituting amino acids at each position of the target p53^R175H peptide (HMTEVVRHC; SEQ ID NO: 1) with each of the remaining 19 common amino acids. T2 cells loaded with each of the 171 variant peptides were then used to assess T-cell activation by measuring IFN-g release following incubation with H2-scDb (Fig. 18D). In congruence with the X-ray structural analysis, any changes in P8, where the mutant histidine residue lies, and any change in P7, which is encased with P8 by the CDR loops, abolished recognition of the peptide. Importantly, peptides with substitutions at these positions retained their ability to bind to HLA-A*02:01 (Fig. 34), but not to the H2-scDb. Other non-anchor residues at positions 3-6 also highly favored the parental amino acids present in the target peptide. This recognition pattern is illustrated as a Seq2Logo graph (Fig. 18E).

Next, a nonamer binding motif, x-[AILMVNQTC]-[ST]-[DE]-[IV]-[IMVST]-R-H- [AILVGHSTYC] (SEQ ID NO: 197), was generated using 20% target peptide reactivity as a cutoff for permissive amino acids at each position (Fig. 35). A search of this motif in the UnitProtPK human protein database using ScanProsite yielded 3 homologous peptides from STAT2 (PLTEIIRHY; SEQ ID NO: 185), VP13A (LQSEVIRHY; SEQ ID NO: 186), and ZFP3 (QNSEIIRHI; SEQ ID NO: 187) (Table 9). None of these 3 peptides were predicted to be potent binders of HLA-A*02:01 by NetMHCpan 4.0 (% rank all >2.0) and had lower predicted binding affinity than the parental p53^R175H peptide (Table 9). However, to experimentally exclude the possibility of cross-reactivity, T2 cells were pulsed with each of these peptides. H2-scDb activated T cells only in the presence of T2 pulsed with the p53^R175H peptide (Fig. 18F). Additionally, COS-7 cells were co-transfected with expression plasmids for HLA-A*02:01 and STAT2 or ZFP3; full-length VP13A was not tested due to its large size (3000 aa). Again, no T-cell activation was detected in the co-culture assay with COS-7 cells expressing the two proteins containing the candidate cross-reactive peptides (Fig. 36).

Table 9. Putative cross-reactive peptides identified through peptide scanning. A binding motif of H2-scDb was determined by positional scanning of the p53^R175H HMTEVVRHC (SEQ ID NO: 1) peptide. Three peptides that conformed to this motif in the UniProtPK human protein database were identified using ScanProsite. NetMHCPan4.0 was used to predict the binding affinity and % rank of these peptides to HLA-A*02:01. The parental p53^R175H target peptide is listed for comparison.

Antitumor activity of the H2-scDb in vivo To determine whether H2-scDb could control tumor growth in vivo , KMS26 cells were engrafted into NOD-SCTD-//?/ (NSG) mice through intravenous injection, establishing widespread, actively growing cancer cells throughout the body. Two models were used to assess the effects of the H2-scDb in combination with human T cells. In an early treatment model, mice were randomized based on luminescence quantification of tumor burden and H2-scDb was subsequently administered through intraperitoneal infusion pumps at 0.3 mg/kg/day, starting one day after tumor inoculation. An irrelevant isotype scDb was administered in parallel as control. H2-scDb markedly suppressed the growth of parental KMS26 cells (Fig. 19A). In contrast, the H2-scDb had no effect on KMS26 cells in which the TP53 gene had been disrupted using CRISPR (Fig. 19B). In the second model, mice were randomized 6 days after tumor inoculation. The H2-scDb was administered at two doses (0.15 and 0.3 mg/kg/day). Both doses resulted in major tumor regressions and were well tolerated as assessed by the absence of significant changes in body weight (Fig. 19C, Fig. 37).

Taken together, these results demonstrate that — can be used to target p53, the most common mutation of the most commonly mutated tumor suppressor gene in human cancers, and, as such, can specifically target cancer cells harboring the mutation.

Materials and Methods

Cell lines and primary T cells

COS-7, T2 (174 x CEM.T2), Raji, HH, AU565, SK-BR-3, KLE, HCT116, SW480, NCI-H441, Saos-2, and CCRF-CEM cells were purchased from American Type Culture Collection (ATCC, Manassas, VA). KMS26, TYK-nu, and HuCCTl were purchased from Japanese Collection of Research Bioresources Cell Bank (JCRB, Osaka, Japan). SigM5 was obtained from DSMZ (Braunschweig, Germany). HEK293FT was obtained from Invitrogen (Thermo Fisher Scientific, Waltham, MA). T2, Raji, Jurkat, HH, AU565, NCI-H441, TOV- 112D, CCRF-CEM, KMS26, TYK-nu, and HuCCTl were cultured in RPMI-1640 (ATCC, 30-2001) with 10% FBS (GE Healthcare, SH30070.03) and 1% Penicillin-Streptomycin (Thermo Fisher Scientific, 15140163). COS-7, SK-BR-3, HCT116, SW480, and Saos-2 were culture in McCoy’s 5 A modified media (Thermo Fisher Scientific, 16600108) with 10% FBS and 1% Penicillin-Streptomycin. SigM5 was cultured in IMDM (Thermo Fisher Scientific, 12440061) with 20% FBS and 1% Penicillin-Streptomycin. HEK293FT was cultured in DMEM (high glucose, pyruvate, Thermo Fisher Scientific, 11995065) with 10% FBS, additional 2 mM GlutaMAX (Thermo Fisher Scientific, 35050061), O.lmM MEM non- essential amino acids (Thermo Fisher Scientific, 11140050), 1% Penicillin-Streptomycin, and 500 pg/mL Geneticin (Thermo Fisher Scientific, 10131027). PBMCs were isolated from leukapheresis samples (Stem Cell Technologies, Vancouver, BC) by standard density gradient centrifugation with Ficoll Paque Plus (GE Healthcare, 17-1440-03). T cells were expanded from PBMCs with addition of the anti-human CD3 antibody (OKT3, BioLegend, San Diego, 317347) at 15 ng/mL for three days. T cells were cultured in RPMI-1640 with 10% FBS, 1% Penicillin-Streptomycin, 100 IU/mL recombinant human IL-2 (aldesleukin, Prometheus Therapeutics and Diagnostics, San Diego, CA), and 5 ng/mL recombinant human IL-7 (BioLegend, 581908). All cells were grown at 37°C in 5% CO2 with humidification.

Detection of neoantigen peptide

HLA-A*02:01-restricted p53^R175H peptide was directly detected and quantified in COS-7 cells transfected with HLA-A*02:01 and p53R175H and in human cancer cell lines expressing HLA-A*02:01 and p53^R175H through MANA-SRM. In particular, the dual reduction approach described in MANA-SRM was critical for this detection because a cystine and a methionine coexist in the p53^R175H peptide. 100 femtomole heavy-isotope labeled peptide HMTEVVRHC (SEQ ID NO: 1; New England Peptide Inc, Gardner, MA) were spiked into each sample before the assay. The MANA-SRM assays were performed at Complete Omics (Baltimore, Maryland).

Phage display library construction

The scFv-bearing phage library was described elsewhere (see, e.g., Miller et al., J Biol. Chem. 294:19322-19334 (2019); and Skora et al., Proc. Natl. Acad. Sci. U. S. A.

112:9967-9972 (2015)). Oligonucleotides were synthesized by GeneArt (Thermo Fisher Scientific) using trinucleotide mutagenesis (TRIM) technology to diversify complementarity determining region (CDR)-L2, CDR-L3, CDR-H1, CDR-H2, and CDR-H3. A FLAG (DYKDDDDK; SEQ ID NO: 190) epitope tag was placed immediately downstream of the scFv, which was followed in frame by the full-length Ml 3 pill coat protein sequence. The total number of transformants obtained was determined to be 3.6xl0¹⁰.

Peptides and monomers

All peptides were synthesized at a purity of >90% by Peptide 2.0 (Chantilly, VA) or ELIM Biopharm (Hayward, CA), except for the positional scanning library, where crude peptides were used. Peptides were resuspended in dimethylformamide at 10 mg/mL and stored at -20°C. Peptide-HLA monomers were synthesized by refolding recombinant HLA with peptide and b2 microglobulin (b2M), purified by gel-filtration, and biotinylated (Fred Hutchinson Immune Monitoring Lab, Seattle, WA). Monomers were confirmed to be folded prior to selection by performing an ELISA using W6/32 antibody (BioLegend, 311402), which recognizes only folded HLA.

Selection of mutant pHLA specific phage clone

Phage clones bearing scFvs specific to p53^R175H/HLA-A*02:01 pHLA were identified using an approach described elsewhere (Skora et al., Proc. Natl. Acad. Sci. U. S. A.

112:9967-9972 (2015)). Biotinylated HLA-A*02:01 pHLA monomer complexes were conjugated to 25 pL of M-280 streptavidin magnetic Dynabeads (Thermo Fisher Scientific,

11206D) per 1 pg of pHLA. During the enrichment phase (Round 1), approximately 4xlO^l3 phage, representing -1000-fold coverage of the library, were negatively selected with a mixture of unconjugated Dynabeads and free streptavidin protein (RayBiotech, Norcross,

GA, 228-11469). After negative selection, supernatant containing unbound phage was transferred for positive selection using 1 pg of p53^R175H/HLA-A*02:01 pHLA. Beads were then washed and phage was eluted to infect mid-log-phage SS320 bacteria, with the addition of M13K07 helper phage (MOI of 4). Bacteria were then grown overnight at 30°C for phage production and the phage was precipitated the next morning with PEG/NaCl.

During the selection phase (Rounds 2-5), phage from the previous round was subjected to two stages of negative selection: 1) against cell lines without p53^R175H/HLA- A*02:01 (RPMI-6666, Jurkat, Raji, SigM5, HH, T2, andNCI-H441) and 2) against p53^WT/HLA-A*02:01 pHLA, unrelated HLA-A*02:01 pHLA, and free streptavidin. For negative selection using cell lines, phage was incubated with a total number of 5xl0⁶ - lxlO⁷ of cells at 4°C overnight. After negative selection, beads were isolated and unbound phage was transferred for positive selection by incubating with 1 pg (Round 2), 0.5 pg (Round 3), or 0.25 pg (Round 4, 5) of p53^R175H/HLA-A*02:01 pHLA. Phage was then eluted and amplified by infecting SS320 as described above.

After five rounds of selection, SS320 cells were infected with a limiting dilution of the enriched phage. A total of 190 individual colonies of SS320 were picked and phage DNA was PCR amplified by primers flanking the CDRs (Forward:

GGC CAT GGC AGAT ATTC AG A (SEQ ID NO: 198), Reverse:

CCGGGCCTTTATCATCATC (SEQ ID NO: 199)) using Q5 Hot Start High-Fidelity 2X Master Mix (New England BioLabs, M0494L) and Sanger sequenced by GENEWIZ (South Plainfield, NJ). Sequences flanking the CDRs were trimmed using DNA Baser Sequence Assembler v4 (Arges, Romania) and the sequences spanning the CDRs were clustered using the CD-HIT Suite. Colonies containing unique phage clones were selected and grown overnight in 400 pL of media in deep 96-well plates (Thermo Fisher Scientific, 278743) with the addition of M13K07 helper phage. Bacteria were pelleted the next day and the phage laden supernatants were used for downstream analysis.

Peptide pulsing

For peptide pulsing, T2 cells were washed with serum-free RPMI-1640 media before incubation at 5xl0⁵ - lxlO⁶ cells per mL in serum-free RPMI-1640 containing specified concentration of peptide for 2 hours at 37°C. For experiments assessed by flow cytometry,

10 pg/mL human b2M (ProSpec, East Brunswick, NJ, PRO-337) was added with the peptides and is specified in the figure legend of such experiments.

Flow cytometry

Phage staining of peptide-pulsed T2 cells was performed with 50 pL phage supernatant on ice for 1 hour, followed by staining with 1 pg of rabbit anti -Ml 3 antibody (Novus Biologicals, NB100-1633), and anti-rabbit-PE (BioLegend, 406421). HLA-A*02 staining was performed by staining cells with fluorescently labeled anti-human HLA-A*02 (BB7.2, BioLegend, 343308) or mouse isotype IgG2b, k (BioLegend), 402206). Stained cells were analyzed using an LSRII flow cytometer (Becton Dickinson, Mansfield, MA) or an iQue Screener (IntelliCyt, Albuquerque, NM).

ELISAs

Streptavidin-coated, 96-well plates (R&D Systems, Minneapolis, MN, CP004) were coated with 50 ng of biotinylated HLA-A*02:01 pHLA monomers in 50 pL of blocking buffer (PBS with 0.5% BSA, 2 mM EDTA, and 0.1% sodium azide) at 4°C overnight. Plates were washed with IX TBST (TBS + 0.05% Tween-20) using a BioTek 405 TS plate washer (BioTek, Winooski, VT). Serial dilutions of single chain diabodies (scDbs) were incubated on the plate for 1 hour at RT, washed then incubated with 1 pg/mL recombinant protein L (Thermo Fisher Scientific, 77679) for 1 hour at RT, washed, then incubated with anti-protein L HRP (1:10000, Abeam, ab63506) for 1 hour atRT. Plates were washed, 50 pL of 3, 3', 5, 5'- Tetramethylbenzidine (TMB) substrate (BioLegend, San Diego, CA, 4211101) was added to each well, and the reaction was quenched with 50m1 2N sulfuric acid (Thermo Fisher Scientific). Absorbance at 450 nm was measured with a Synergy HI Multi-Mode Reader (BioTek).

Bispecific antibody production

Single chain diabodies (scDbs) were produced by cloning gBlocks (IDT, Coralville, Iowa) encoding each of the variants in the format of, from N- to C-terminus: IL-2 signal sequence, anti-pHLA variable light chain (VL), GGGGS linker (SEQ ID NO:200), anti-CD3 variable heavy chain (VH), (GGGGS)3 linker (SEQ ID NO:201), anti-CD3 VL, GGGGS linker (SEQ ID NO:200), anti-pHLA VH, and 6 x HIS tag into linearized pcDNA3.4 vector (Thermo Fisher Scientific, A14697). The proteins were expressed by the Eukaryotic Tissue Culture Core Facility of Johns Hopkins University. Briefly, 1 mg of plasmid DNA was transfected with PEI at a ratio of 1 :3 into 1 L of FreeStyle 293-F cells at a density of 2- 2.5xl0⁶ cells per mL and incubated at 37°C. Five days after transfection, culture media was collected and filtered through a 0.22-pm unit. The scDbs were purified using HisPur Ni- NTA Resin (Thermo Fisher Scientific, 88222) and desalted into PBS pH 7.4 or 20 mM Tris pH 9.0, 150 mM NaCl using 7k MWCO Zeba Spin desalting columns (Thermo Fisher Scientific, 89890). Proteins were quantified using a 4-15% Mini-PROTEAN TGX gel (Bio- Rad, Hercules, CA, 4568085) and/or nanodrop (Thermo Fisher Scientific). Proteins were stored at 4°C for short term storage or snap frozen with the addition of 7% glycerol and stored at -80°C for long term storage. Alternatively, the scDb protein was produced by GeneArt (Thermo Fisher Scientific) in Expi293s, purified with a HisTrap column (GE Healthcare, 17-5255-01) followed by size exclusion chromatography with a HiLoad Superdex 20026/600 column (GE Healthcare, 28989335).

Surface plasmon resonance affinity measurements of p53^R175H/HLA-A *02:01 to H2-scDb

Biotinylated p53^R175H/HLA-A*02:01, p53^WT/HLA-A*02:01, and H2-scDb binding experiments were performed at 25°C using a Biacore T200 SPR instrument (GE Healthcare). Approximately 100-110 response units (RU) of biotinylated p53^R175H/HLA-A*02:01 and p53^WT/HLA-A*02:01 were captured in flow cells (Fc) 2 and 4, respectively, using a streptavidin chip. Single-cycle kinetics were performed by injecting increasing concentrations (3, 12, 50, 200 to 800 nM) of purified clone H2-scDb flowed over Fc 1-4. Binding responses for kinetic analysis were both blank- and reference-subtracted. Both binding curves were fit with a 1 : 1 binding model using Biacore Insight evaluation software.

CRISPR-mediated knockout of TP 53

The Alt-R CRISPR system (IDT) was used to knock out the TP53 gene from KMS26, TYK-nu, and KLE cell lines. CRISPR Cas9 crRNAs targeting IP 33 exon 3 (p53-5: CCCCGGACGATATTGAACAA (SEQ ID NO: 191) or p53-6:

CCCCTTGCCGTCCCAAGCAA (SEQ ID NO: 202)) as well as CRISPR-Cas9 tracrRNA were resuspended at 100 mM with Nuclease-Free Duplex Buffer. The crRNAs and tracrRNA were duplexed at a 1:1 molar ratio for 5 minutes at 95°C. The duplexed RNA was allowed to cool to room temperature prior to mixing with Cas9 Nuclease at a 1.2: 1 molar ratio for 15 minutes. A total of 40 pmols of the Cas9 RNP complexed with TP53 gRNA were mixed with 200,000 cells in 20 pL of OptiMEM. This mixture was loaded into a 0.1 cm cuvette (Bio-Rad, 1652089) and electroporated at 120 V and 16 ms using an ECM 2001 (BTX, Holliston, MA). Cells were immediately transferred to complete growth medium and cultured for 7 days. Single cell clones were established by limiting dilution and genomic DNA was harvested using a Quick-DNA 96 Kit (Zymo Research, Irvine, CA, D3012). A region flanking the CRISPR cut site was PCR amplified (forward primer: GCTGCCCTGGTAGGTTTTCT (SEQ ID NO:203), reverse primer: GAGACCTGTGGGAAGCGAAA (SEQ ID NO:204)) and Sanger Sequenced to select for clones with the desired TP53 status.

Immunoblotting analysis

Cells were lysed in cold RIPA buffer (Thermo Fisher Scientific, 89901) supplemented with protease inhibitor cocktail (Thermo Fisher Scientific, 87785). Protein concentration was determined using a BCA assay (Thermo Fisher Scientific, 23227). Equal amounts of total protein (20-50 pg) were loaded in each lane of a 4-15% Mini-PROTEAN TGX gel (Bio-Rad, 4568085) and transferred to polyvinylidene difluoride membranes after electrophoresis. The membranes were incubated with appropriate primary antibodies (p53 [DO-1], 1:1000, Santa Cruz, sc-126; STAT2, 1:1000, Thermo Fisher Scientific, 44-362G; ZFP3, 1:1000, Thermo Fisher Scientific, PA5-62726; b-actin [13E5], 1:1000, Cell Signaling Technology, 5125S; b-actin [8H10D10], 1:1000, Cell Signaling Technology, 3700S) and species-specific HRP-conjugated secondary antibodies (1:5000-10000). Signal was detected by a ChemiDoc MP chemiluminescence system (Bio-Rad).

Transfection of cell lines gBlocks (IDT) encoding HLA and target proteins were cloned into pcDNA3.1 or pcDNA3.4 vectors (Thermo Fisher Scientific, V79020, A14697). COS-7, Saos-2, and HEK293FT cells were transfected at 70-80% confluency using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015) and incubated at 37°C overnight. A total of 15 pg and 30 pg plasmid (1 : 1 ratio of HLA plasmid/target protein plasmid in co-transfections) was used for T25 and T75 flasks, respectively.

Viral transduction of cell lines

HLA-A*02:01-encoding retrovirus was produced using the MSCV retroviral expression system (Clontech, Mountain View, CA, 634401). In brief, a gBlock encoding HLA-A*02:01-T2A-GFP (IDT) was cloned into the pMSCVpuro retroviral vector by HiFi DNA assembly (New England Biolabs, Ipswich, MA, E2621L). The pMSCVpuro-HLA- A*02:01-T2A-GFP plasmid was then co-transfected with a pVSV-G envelope vector into the GP2-293 packaging cell line. Viral supernatant was harvested 48 hours after transfection and concentrated 20-fold using Retro-X Concentrator (Clontech, 631456). RediFect Red-Fluc- GFP lentivirus particles (Perkin Elmer, Waltham, MA, CLS960003) was used for generating luciferase-expressing cell lines.

For transduction, non-tissue culture-treated 48-well plates were coated with 200 pL of 10 pg/mL RetroNectin (Clontech, T100B) per well overnight at 4°C and blocked with 10% FBS for 1 hour at RT. Viral particles and 2xl0⁵ target cells were added to each well in a total volume of 500 pL cell culture media and spun at 2000 x g for 1 hour then incubated at 37°C. Selection with 1 pg/mL puromycin (Thermo Fisher Scientific, A1113803) began three days later. Transduced cells were sorted based on presence of GFP using FACSAria Fusion (BD Biosciences, San Jose, CA) 10-14 days after transduction. in vitro scDb co-incubation assays

To each well of a 96-well flat-bottom plate, the following components were combined in a final volume of 100 pL RPMI-1640 with 10% FBS, 1% Penicillin- Streptomycin, and 100 IU/mL IL-2: bispecific antibody diluted to the specified concentration, 5xl0⁴ human T cells, and l-5xl0⁴ target cells (COS-7, T2, or other tumor cell lines). The effector to target cell ratio is specified in the figure legend for each experiment. The co-culture plate was incubated for 20 hours at 37°C and conditioned media was assayed for cytokine and cytotoxic granule protein secretion using the Human IFN-g Quantikine Kit (R&D Systems, Minneapolis, MN, SIF50), Human IFN-g Flex Set Cytometric Bead Array (BD, 558269), or the MILLIPLEX Luminex assays (Millipore Sigma,

HS T CM AG28 SPMX 13 , HCD8MAG-15K) read on the Bioplex 200 platform (Bio-Rad). Cytotoxicity was assayed by CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, G7571) or Bio-Glo Luciferase Assay (Promega, G7941) per manufacturer’s instructions. For CellTiter-Glo assays, percent cytotoxicity was calculated by subtracting the luminescence signal from the average of the T cell only wells and normalizing to the no scDb condition: 1 - (scDb well - T cell only)/(no scDb well - T cell only) x 100. For Bio-Glo assays, percent cytotoxicity was calculated by normalizing luminescence signal to the no scDb condition: 1 - (scDb well)/(no scDb well) x 100.

Real-time live-cell imaging

A total of lxlO⁴ CellTracker Green CMFDA (Thermo Fisher Scientific, C7025)- labeled target cells were plated in each well of a 96-well flat bottom plate and allowed to attach for 8 hours before adding T cells and scDb at the indicated E:T ratio and concentrations, respectively. Each condition was plated in triplicate. Plates were imaged every 3 hours using the IncuCyte ZOOM Live-Cell analysis system (Essen Bioscience, Ann Arbor, MI) for a total of 60 hours. Four images per well at 10X zoom were collected at each time point. Cell confluence in each well was quantified using the phase contrast channel.

Expression, purification and refolding of p53^R175H/HLA-A *02:01

Plasmids for HLA-A*02:01 and b2M were received from the NIH Tetramer Facility (Atlanta, GA) and separately transformed into BL21(DE3) cells. Each was expressed in inclusion bodies using auto-induction media as described elsewhere (Skora et al., Proc. Natl. Acad. Sci. U S. A. 112:9967-9972 (2015); Martayan et al., The Journal of Immunology 182:3609-3617 (2009); and Huang et al., Bioinformatics 26:680-682 (2010)). Purification of the HLA-A*02:01 and b2M inclusion bodies was achieved with a series of detergent washes followed by solubilization with 8 M urea. Refolding of the HLA-A*02:01, b2M, and mutant p53^R175H peptide was performed as described elsewhere (Myszka, J. Mol. Recognit. 12:279- 284 (1999)). Briefly, HLA-A*02:01 and b2M inclusion bodies were combined in a refolding buffer containing 100 mM Tris pH 8.3, 400 mM L-arginine, 2 mM EDTA, 5 mM reduced glutathione, 0.5 mM oxidized glutathione, 2 mM PMSF, and 30 mg of the mutant p53^R175H peptide (amino acid 168-176, HMTEVVRHC (SEQ ID NO:l)) dissolved in 1 mL of DMSO. The resultant solution was stirred at 4°C for 2 days, with two further additions of HLA- A*02:01 inclusion bodies on day 2, concentrated to 10 mL and purified by size exclusion chromatography on a HiLoad 26/60 Superdex 75 Prep grade column (GE Healthcare, 28989334). For incubation with the H2-Fab, purified pHLA-A*02:01 was concentrated to ~l-3 mg/mL and stored at -80°C until use.

Production of the H2-Fab antibody fragment

The light chain (LC) and heavy chain (HC) variable region sequences of H2 scFv were grafted, linked with the respective constant region sequences of human IgGl and separately cloned into a pcDNA3.4 vector (Thermo Fisher Scientific, A14697). Both chains were preceded by a mouse IgKVIII signal peptide. Before large-scale expression of full- length antibody, optimization of the LC:HC DNA ratio for transfection was performed to determine optimal recombinant protein yields. For a 1 L expression, a total of 50 pg of purified plasmids (1 : 1 LC:HC ratio) were transfected with PEI at a ratio of 1 :3 into Freestyle 293-F cells at a density of 2-2.5xl0⁶ cells per mL and incubated at 37°C for 7 days. The media was harvested via centrifugation, filtered through a 0.22-pm unit and the full-length antibody was purified via protein A affinity chromatography on a HiTrap Mab Select™ SuRe™ column (GE Healthcare, 29-0491-04). Full-length antibody was eluted using a linear gradient of 0-100 mM sodium citrate, pH 3.5. The protein A fractions containing pure H2 antibody were pooled, quantified by SDS-PAGE gel electrophoresis and dialyzed into 20 mM sodium phosphate buffer, pH 7.0, 10 mM EDTA. For generation of H2-Fab fragments, ~l-3 mg of full-length antibody was mixed with 0.5 mL of a 50 % Immobilized Papain slurry (Thermo Fisher Scientific, 20341) pre-activated with digestion buffer (20 mM sodium phosphate buffer, pH 7.0, 10 mM EDTA) containing 20 mM cysteine-HCl. The mixture was incubated at 37°C overnight with constant shaking at 200 rpm. The H2 antibody digest was separated from the immobilized resin by a gravity resin separator and washed with 10 mM Tris-HCl, pH 7.5. Newly generated H2-Fab fragments were further purified by cation-exchange chromatography using a Mono-S column (GE Healthcare, 17516801) and eluted using a linear gradient of 0-500 mM NaCl.

The H2-Fab fragments were concentrated, mixed with equimolar p53^R175H/HLA- A*02:01 and incubated at 4°C overnight. The H2-Fab-p53^R175H/HLA-A*02:01 mixture was evaluated by size exclusion chromatography on a Superdex™ 200 Increase 10/300 column (GE Healthcare, 28990944). The fractions of- 98% pure pHLA-A*02:01-H2-Fab complex were pooled, concentrated to 12.6 mg/mL and exchanged into a buffer containing 25 mM HEPES, pH 7.0, 200 mM NaCl.

Crystallization, data collection and structure determination

Crystals of the ternary complex H2-Fab-p53^R175H/HLA-A*02:01 were grown in hanging drop by vapor diffusion in drops set up with a TTP mosquito robot with a reservoir solution of 0.2 M ammonium chloride and 20% (w/v) PEG 3350 MME. Crystals were flash- frozen in mother liquor. Data were collected at National Synchrotron Light Source-II at beamlines 17-ID-1(AMX) on a DECTRIS Eiger X 9M detector and 17-ID-2 (FMX) on a DECTRIS Eiger X 16M detector. Datasets were indexed, integrated and scaled using fastdp, XDS, and aimless. Monoclinic crystals of H2-Fab-p53^R175H/HLA-A*02:01 diffracted to 3.5 A. The structure for the H2-Fab-p53^R175H/HLA-A*02:01 complex was determined by molecular replacement with PHASER using PDB ID 604Y (77) and 6UJ9 as the search models. The data was refined to a final resolution of 3.5 A using iterative rounds of refinement with REFMAC5 and manual rebuilding in Coot. Structures were validated using Coot and PDB Deposition tools. The model has 94.11% of the residues in preferred and allowed regions according to Ramachandran statistics (Table 7). Figures were rendered in PyMOL (v2.2.3, Schrodinger, LLC, New York, NY). Buried areas were calculated with PDBePISA. The angle that determines the relative orientation between the pHLA and the Fab/TCR was calculated by the dot product of the vector defined by the position of the alpha carbon of the P1-P9 of the peptide and the vector defined by the disulfide bonds in the VH and VL domains.

Mouse xenograft model

Female NOD . ( 'g-Prkdc^scldll2rf^{m 1 ,l/V}/SzJ (NSG) mice at 6-10 weeks were acquired from the Jackson Laboratory (Bar Harbor, Maine, 005557) and treated in compliance with the institutional Animal Care and Use Committee approved protocol. In the early treatment model, mice were inoculated intravenously with lxlO⁶ luciferase-expressing KMS26 or KMS26-ZP53 KO cells and lxlO⁷ in vitro expanded human T cells via lateral tail vein injection on day 0. On day 1, mice were randomized based on luminescence quantification using the IVIS imaging system and Living Image software (Perkin Elmer) to ensure similar pretreatment tumor burden. After randomization, two-week micro-osmotic pumps (ALZET, Cupertino, CA, 1002) filled with H2-scDb or isotype control scDb (scFv against an irrelevant pHLA linked with UCHTlscFv) that had been primed in 1 mL PBS overnight at 37°C were placed intraperitoneally using sterile surgical technique. Tumor growth was serially monitored by bioluminescent imaging. In the established tumor model, mice were inoculated with 3.5xl0⁵ luciferase-expressing KMS26 cells and lxlO⁷ human T cells via lateral tail vein injection on day 0. On day 6, H2-scDb or isotype control scDb was administered similarly as in the early treatment model.

Statistical analysis

Data are presented as means ± SD. Statistical analyses were carried out using specific tests indicated in the figure legend. A P value of < 0.05 was used to denote statistical significance. All analyses were performed using Prism version 8.0 (GraphPad, San Diego, CA).

Example 3: Bispecific Antibodies Targeting Mutant RAS Neoantigens

Mutations in the RAS oncogenes occur in multiple cancers and new approaches to target these mutations have been the subject of intense research for decades. Most of these efforts have been focused on conventional small molecule drugs rather than antibody -based therapies because the RAS proteins are intracellular. This Example identifies scFvs specific for peptides derived from two recurrent RAS mutations, G12V and Q61H/L/R, which are presented on cancer cells in the context of two common human leukocyte antigen alleles, HLA-A3 and HLA-A1, respectively. The scFvs did not recognize the peptides derived from the wildtype (WT) form of RAS proteins or other related peptides. Given their extremely low antigen density on cancer cells, a very sensitive immunotherapeutic agent was developed to kill cells harboring RAS gene mutations. Single chain diabodies (scDbs) specific for peptides derived from G12V or Q61H/L/R were capable of inducing T cell activation and killing of target cancer cells expressing endogenous levels of the mutant RAS proteins and cognate HLA alleles.

Results

MANAs derived from clinically relevant RAS gene mutations

In silico predictions suggested that the 10-mer peptide from codons 5 to 14 (KLVVVGAVGV, “G12V[5-14]”; SEQ ID NO:209), in which the underlined valine residue (V) represents the G12V mutation, would bind to HLA-A*02:01 (henceforth referred to as HLA-A2) (see, e.g., Skora et al., Proc Natl Acad Sci USA 112:9967-9972 (2015); and Andreatta, Bioinformatics 32:511-517 (2016)). A MANAbody (called D10) that bound to this pHLA-A2 complex and did not bind to the wildtype (WT) counterpart, with selective binding demonstrated using pHLA complexes attached to artificial surfaces as well as using cells pulsed with mutant peptides, was previously reported (Skora et al., Proc Natl Acad Sci USA 112:9967-9972 (2015)). However, it could not be demonstrated that the D10 MANAbody could bind to cells that expressed endogenous levels of the mutant KRAS gene or even to cells that overexpressed an exogenous mutant KRAS gene. It was hypothesized that despite the predictions of the in silico algorithms, the KRAS protein was not processed to the G12V[5-14] peptide and transported to the cell surface. There is indeed precedent for such in silico predictions to be inaccurate (Schmidt et al., J Biol Chem 292: 11840-11849 (2017)). To evaluate this possibility, a highly-sensitive mass spectrometry (MS)-based approach (MANA-SRM) was developed to analyze HLA-bound peptides. Human HLA-A2 and full-length human KRAS containing G12V were expressed in SV40 virus-immortalized monkey kidney COS-7 cells and HLA-bound peptides were immunopurified. MS analysis of eluted peptides showed that the G12V[5-14] peptide could not be detected, even when the mutant KRAS gene was overexpressed (Table 10). It was concluded that this peptide was not processed and presented on HLA-A2+ cells harboring a KRAS G12V mutant gene.

Table 10. MANA-SRM quantification. Mutation-associated Neoantigen (MANA) peptides per cell were quantified using the mass-spectrometry based method MANA-SRM. COS-7 cells were co-transfected with plasmids encoding HLA alleles and full-length KRAS variants. To assess endogenous expression of MANAs, cell lines bearing HLA alleles and RAS variants of interest were analyzed. Thus, attention was to two peptides, a 9-mer (VVGAVGVGK; SEQ ID NO:206) from codons 8 to 16 (“G12V[8-16]”) and a 10-mer (VVVGAVGVGK; SEQ ID NO:205) from codons 7 to 16 (“G12V[7-16]”), containing the RAS G12V mutation that were predicted by NetMHCv4.0 to bind HLA-A*03:01 (henceforth referred to as HLA-A3) with high affinity (Andreatta, Bioinformatics 32:511-517 (2016)). HLA-A3 is one of the most common HLA-A alleles. MANA-SRM was used to confirm the in silico predictions that these peptide complexes would be displayed on the cell surface before screening for MANAbodies. Human HLA-A3 and full-length human KRAS G12V were expressed in COS-7 cells, then immunopurified with an anti-human HLA antibody. MS analysis of peptides eluted from the captured peptide-HLA complexes detected the G12V[7-16] peptide at 102 copies per cell and the G12V[8-16] peptide at considerably lower levels (24 copies per cell) (Fig. 45A, Table 10). Importantly, when COS-7 cells were transfected with HLA-A2 and KRAS G12V as a negative control, neither of these peptides were detected (Table 10). MANA-SRM analysis was then performed on two human cancer cell lines, the lung adenocarcinoma line NCI-H441 and the pancreatic ductal adenocarcinoma line CFPAC-1, both expressing endogenous levels of HLA-A3 and harboring the KRAS G12V mutation.

The CFPAC-1 and NCI-H441 lines presented on average 3 and 9 copies per cell, respectively, of the G12V[7-16] peptide (Fig. 45, B and C, Table 10), and the G12V[8-16] peptide was not detected. Given the higher abundance of the 10-mer G12V[7-16]

(henceforth referred to as “G12V”) peptide, efforts were focused on targeting this MANA.

Another commonly mutated residue in RAS genes is the glutamine at codon 61. RAS codon 61 mutations are found in a wide variety of cancers, including melanomas, multiple myelomas, thyroid, and bladder cancers. Using NetMHCv4.0, it was predicted that a 10-mer RAS peptide (codons 55 to 64) could bind with high affinity to HLA-A*01 :01 (henceforth referred to as HLA-A1), another common HLA-A allele (Maiers et al., Hum Immunol 68:779-788 (2007)). This peptide (ILDTAGQEEY; SEQ ID NO: 136) as well as the flanking amino acids are conserved across all three RAS proteins. Using MANA-SRM, cell surface expression of 10-mer peptides from codons 55 to 64 containing the Q61H, Q61L, and Q61R mutations (ILDTAGHEEY (SEQ ID NO:2), ILDTAGLEEY (SEQ ID NO:3), and ILDTAGREEY (SEQ ID NO:4), respectively) were previously evaluated, and in the transfected COS-7 overexpression system, an average of 583, 512, and 127 copies of the Q61H, Q61L, and Q61R peptides per cell were found (Wang et al., Cancer Immunol Res 7:1748-1754 (2019)). These peptides were also identified to be presented on cell lines expressing endogenous levels of Q61 mutant RAS proteins, notably with four copies of the Q61L peptide per cell found in the acute promyelocytic leukemia cell line HL-60 (Table 10).

Together, these data show that G12V and Q61 mutant RAS proteins can be processed into peptides that are presented on the surface of cancer cell lines, albeit at antigen densities below what is considered the minimum required for recognition by conventional antibody- based immunotherapeutic agents.

Identification of scFv-expressing phage clones targeting HLA-restricted RAS mutant peptides

To develop MANAbodies targeting the above-named RAS peptides, a second phage library displaying scFvs was built. This library was designed based on principles described elsewhere (Skora et al., Proc Natl Acad Sci USA 112:9967-9972 (2015)), but with important modifications. In particular, precursor library DNA was synthesized using trinucleotide mutagenesis (TRIM) technology, permitting fine tuning of the amino acid diversity at particular codons considered most critical for antigen binding. Diversity was introduced in five of the six complementarity-determining regions (CDRs), with the most amino acid diversity as well as length diversity incorporated into the third CDR of the heavy chain (CDR-H3) (Fig. 46, A and B). The completed library was estimated to contain ~3.6 x 10¹⁰ unique clones. As a quality control step, a portion of the library was subjected to massively parallel sequencing of the heavy chain CDR3 region, demonstrating that the expected and actual amino acid diversity were in excellent alignment (Fig. 46C).

For the RAS G12V HLA-A3 MANA target, HLA-A3 and beta-2-microglobulin were folded together with the chemically synthesized G12V peptide or its WT [7-16] (G12WT) counterpart to form pHLA complexes (Table 11). These pHLA-A3 were then used to screen the phage display library described above. The screening process consisted of four to six rounds of phage selection, similar to the scheme described previously, with key differences outlined in the Materials and Methods. Over the course of this selection, phage were negatively selected against soluble streptavidin, streptavidin magnetic beads, denatured HLA-A3, unrelated pHLA-A3, and G12WT pHLA-A3 and positively selected against the G12V pHLA-A3. Following the screening process, individual phage clones were amplified and subjected to enzyme-linked immunosorbent assay (ELISA) to assess their binding specificity (Fig. 47A). While multiple clones initially appeared to be promising on ELISAs with plaste-bound pHLA complexes, one phage clone (V2) stood out upon subsequent flow cytometry assays on cells displaying these complexes (Fig. 47B, Table 12). The V2 phage had substantially greater binding to G12V pHLA-A3 compared to other pHLA-A3, including those formed with HLA-A3 folded with the G12WT peptide, the G12C or G12D variant, or an unrelated CTNNB peptide (Fig. 47C, Table 11). V2 phage also failed to detectably interact with HLA-A3 folded with the G12V[8-16] or G12WT[8-16] peptides (Fig. 47C, Table 11).

Table 11. Peptides for pHLA complexes. Peptides used for pHLA complexes were predicted to bind to the HLA allele of interest using NetMHCv4.0. Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 12. scFv sequences. Amino Acid (AA) sequences of anti-CD3 seFvs used for bi specific antibody construction. Affinities of select anti-CD3 clones were gathered from the literature.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

The scFv component of the phage was then expressed and purified in bacteria for further characterization. The recombinant V2 scFv retained the same binding profile as its phage counterpart (Fig. 47D), with only minimal binding to the WT pHLA complex at the highest concentration tested (Fig. 38 A). TAP-deficient T2 cells modified to express HLA- A3 (T2A3) were pulsed with the G12V peptide or a variety of control peptides, and the cells were assessed for binding to V2 scFv or HLA- A3 -specific monoclonal antibodies. Flow cytometric analysis showed remarkably specific binding of V2 scFv to G12V peptide pulsed T2A3 cells compared to cells pulsed with the G12WT, G12C, or G12D RAS peptides (Fig. 38B, Fig. 47E). As in the ELISA assay, the V2 scFv was unable to detectably bind to cells pulsed with 9-mer [8-16] peptides, whether mutant or WT. Surface plasmon resonance (SPR) binding analysis of the V2 scFv demonstrated a KD value of 8.7 nM for G12V pHLA- A3, while no appreciable binding to G12WT pHLA-A3 was identified (Fig. 38C).

A similar screening procedure was used for RAS Q61H, Q61L, and Q61R pHLA-Al MANA targets (see Materials and Methods). Using phage staining and T cell-based assays, one phage clone was identified displaying high specificity for each of the three targets: clone HI for Q61H, clone L2 for Q61L, and clone R6 for Q61R (Fig. 48, A to F and Table 12). These clones bound to their cognate mutant pHLA with no detectable binding to the RAS Q61WT pHLA-Al (Fig. 38, D to F). To further assess binding to pHLA complexes on the cell surface, the HLA-A1+ acute myeloid leukemia line SigM5 was pulsed with cognate peptides or controls. All three clones specifically bound to the expected pHLA complexes derived from mutant RAS genes (Fig. 38G). Given the efficient presentation of the Q61L peptide on transfected COS-7 cells (27) and its higher level of binding to peptide-pulsed SigM5 cells, most subsequent experiments focused on Q61L clone L2. SPR analysis of the L2 clone revealed a KD of 65 nM for Q61L pHLA-Al and no appreciable binding to Q61WT pHLA-Al (Fig. 38H).

T cell-engaging bispecific antibodies can recognize mutant RAS-derived pHLA complexes

A variety of T cell-engaging bispecific antibody formats have been developed for targeting T cells to specific ligands. However, there are little data available on whether any of these formats can recognize targets when they are presented on the cell surface at low densities. To inform this point, six bispecific formats were evaluated: diabodies, single chain diabodies (scDbs), bispecific T cell engagers (BiTEs), dual affinity retargeting molecules (DARTs), bivalent scFv-Fcs, and trivalent scFv-Fcs (Fig. 49A). The scFv-Fcs were heterodimerized via the Knob-Into-Hole method. The V2 scFv was used as the pHLA- targeting moiety and different anti-CD3 antibodies were used for engaging T cells (Tables 12 and 13). In several of the formats, different configurations of the heavy (“H”) and light (“L”) chains of the V2 and anti-CD3 scFvs were tested (Fig. 49B).

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 13. V2 bispecific antibody sequences. Amino acid sequences of the G12V pHL A- A3 -targeting V2 bispecific antibodies tested. The anti-CD3 clone UCHT1 is referred to as "U" and anti-CD3 clone UCHTl.v9 is referred to as "U2". The bivalent scFv-Fc was the product of co-expression of V2-FcHole and an anti-CD3-FcKnob. The trivalent scFv-Fc was the product of co-expression of V2-FcHole 5 and an V2-FcKnob-anti-CD3.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

In sum, 42 recombinant proteins were expressed in HEK293 cells to identify the most effective format and configuration (Fig. 49, A and B). Binding to G12V pHLA-A3 and recombinant CD3e/5 protein (Fig. 49C) were assessed by ELISAs at various bispecific protein concentrations. While a few formats exhibited weak binding to one or both antigens, most had similar performance characteristics upon ELISA (Fig. 49C). To further compare the formats, T2A3 cells were pulsed with two concentrations of the G12V peptide and co cultured with T cells and each of the V2 bispecific proteins at two concentrations, followed by measurement of IFNy release to assess T cell activation (Fig. 49D). Despite their similar performance in ELISA assays, the ability of the bispecific formats to recognize the G12V peptide at low antigen densities on cells was highly variable. scDbs generally performed better than other formats, particularly at lower concentration of antibodies. Switching the order of the heavy and light variable domains in the analogous proteins ("LHLH" to "HLHL") abolished their ability to activate T cells, despite these formats showing equivalent functionality on ELISA (Fig. 49). Similarly, while the bivalent scFv-Fc and trivalent scFv- Fc performed particularly well in the ELISAs, they consistently performed poorly when assessed in peptide-pulsed cells.

In the initial tests of formats and configurations, five different anti-CD3 scFvs were used. Based on the results indicating that the scDb performed best, seven additional anti- CD3 clones were tested (Tables 12 and 13). A total of twelve V2 scDbs were expressed, purified (Fig. 50A), and then tested with ELISA for their ability to bind the CD3e/5 protein (Fig. 50B). Two target cell lines were used to further assess these V2 scDbs. The first was KRAS G12V and HLA-A3 co-transfected COS-7 and the second was the lung cancer cell line NCI-H441, which expressed endogenous HLA-A3 and mutant KRAS G12V. In a co-culture experiment with the COS-7 overexpression system, ten of the scDbs activated T cells as shown by substantial IFNy release (Fig. 50C). However, with NCI-H441 cells, only the scDbs containing the UCHT1 ("U") and UCHTl.v9 ("U2") anti-CD3 clones activated T cells (Fig. 50D), with the UCHTl-based scDb (Fig. 39A) performing particularly well. Both scDbs (hereafter referred to as V2-U and V2-U2, respectively) (Tables 12 and 13) retained the remarkable specificity of V2 phage toward G12V pHLA-A3, as they failed to interact with pHLA-A3 folded with the other RAS peptides or an unrelated CTNNB peptide (Fig. 39B, Fig. 51 A). They also bound to the CD3e/5 heterodimer (Fig. 2B, Fig. 51 A). Both scDbs could simultaneously interact with G12V pHLA-A3 and CD3e/5 heterodimer, as shown by a “sandwich” ELISA (Fig. 5 IB), suggesting that the proteins were folded and functioned properly. Given all these data on relative expression, ELISA, and cellular reactivity, the V2-U scDb (N-terminus-Lv2-Hu-Lu-Hv2-C-terminus) was chosen as a focus (Fig. 39A) for most further studies, though V2-U2 was also used for select assays.

Based off of this work with the G12V clone V2, the Q61 scFv clones HI, L2, and R6 were grafted into the optimized scDb format to generate Hl-U, L2-U, and R6-U scDbs. All three scDbs retained binding to their cognate mutant-derived Q61 pHLA-Al and recombinant CD3e/5 on ELISA and none exhibited appreciable binding to the Q61WT pHLA or to other control pHLA including an unrelated CMV pHLA (Fig. 39, C to E and fig. 51, C to E, Table 11). This result showed that the format and configuration chosen on the basis of the RAS G12V scFv was generalizable, and applicable to three other scFvs with varying amino acid sequences and independent targets. scDbs recognize cells pulsed with low nanomolar concentrations of exogenous peptides

To assess the minimal concentration of target antigen required for activating T cells, T2A3 cells were pulsed with G12V or G12WT peptides and then co-cultured with healthy donor T cells in the presence of the V2-U scDb. T cells were activated at G12V peptide concentrations as low as 1 nM, as evidenced by åFNy and TNFa secretion (Fig. 40A, Fig. 52A). Furthermore, even at low peptide concentrations, the scDb mediated antigen- dependent lysis of the peptide-pulsed T2A3 cells (Fig. 40B). There was no appreciable background cytokine secretion or cell killing when cells were pulsed with the G12WT peptide. Similarly, specific activation was seen with the V2-U2 scDb (Fig. 40, A and B).

To determine the approximate antigen density on peptide-pulsed T2A3 cells, the cells were stained with the V2 scFv and assessed by flow cytometry. Antigen density was determined using QIFIKIT and Quantibrite Beads, which allowed quantitative determination of the number of cell surface antigenic molecules (Fig. 52B). It was found that cells pulsed with 320 nM of G12V peptide expressed 300 to 800 antigenic molecules per cell. This flow cytometric method did not have the sensitivity required to detect cells pulsed with <320 nM of G12V peptide. However, 300-fold lower concentrations of pulsed G12V peptide could activate T cells (Fig. 40B). Though there may not be a linear relationship between the concentration of the pulsed peptide and the number of peptide antigens displayed on the cell surface, these results suggest that the number of antigenic molecules per cell recognized by the V2-U scDb is far less than 300. An analogous peptide-pulsing experiment was performed with HLA-A3+ monocyte-derived immature dendritic cells (iDCs) to present the peptide and assayed for T cell secretion of IFNy and TNFa. Both V2-U and V2-U2 scDbs were able to activate T cells when the iDCs were pulsed with G12V peptide at low nanomolar range (Fig. 52, C and D).

Similar experiments were performed on the RAS Q61 targeting scDbs using peptide- pulsed SigM5 cells and iDCs. The L2-U scDb elicited T cell activation as shown by IFNy secretion and SigM5 cytotoxicity when the target cells were pulsed with the Q61L peptide at concentrations as low as 1 nM, without cross-reactivity to the Q61WT peptide (Fig. 40, C and D). While the Hl-U and R6-U scDbs also showed no significant Q61WT peptide cross reactivity, it required 100-fold more peptide (100 nM) for them to activate T cells (Fig. 52, E to H). Likewise, the L2-U scDb could activate T cells in the presence of iDCs at lower peptide concentrations than could the Hl-U and R6-U scDbs (Fig. 52, 1 to N). scDbs recognize COS-7 cells overexpressing HLA and mutant RAS genes

COS-7 cells were co-transfected with plasmids encoding HLA-A3 and either mutant or WT full-length KRAS (Fig. 53 A), then co-cultured with T cells. As noted in the first section of the results, the COS-7 cells expressed -100 copies of the GUV pHLA complex per cell. Addition of the V2-U scDb resulted in secretion of both IFNy and TNFa from T cells in a dose-dependent manner only in the presence of COS-7 cells co-transfected with plasmids encoding HLA-A3 and KRAS GUV (Fig. 41A, Fig. 53B). As little as 10 pM of the V2-U scDb could activate T cells in this experiment (Fig. 41 A, Fig. 53B). A similar activation of T cells by transfected COS-7 cells was observed in the presence of 30 pM of the V2-U2 scDb (Fig. 53, C and D). The secretion of both IFNy and TNFa was highly specific as the target cells expressing HLA-A3 and WT, G12C, or G12D KRAS did not trigger secretion of these cytokines (Fig. 41A, Fig. 53, B to D).

Analogous experiments were performed with the mutant RAS Q61 -targeting Hl-U, L2-U, and R6-U scDbs. COS-7 cells were co-transfected with plasmids encoding HLA-A1 and full-length WT or mutant HRAS, KRAS, or NRAS to assess whether each scDb was capable of recognizing the cognate mutant peptides derived from each of the RAS proteins. Each scDb elicited T cell responses highly specific for the COS-7 cells expressing the RAS gene with the Q61 mutation of interest, regardless of the RAS gene assessed (Fig. 41, B to D). Subnanomolar concentrations of the scDbs were sufficient to activate T cells when co cultured with COS-7 cells harboring Q61 mutations in any of the three genes. Moreover, no activation of T cells was observed when co-cultured with COS-7 cells expressing HRAS , KRAS , or NRAS genes that were WT or contained the non-cognate Q61 mutations (Fig. 41, B to D). scDbs activate T cells when exposed to cancer cells harboring endogenous mutant RAS genes

As noted above, the NCI-H441 cancer cell line presents only ~9 copies of mutant- derived KRAS G12V pHLA complexes per cell (Table 10). Despite this extremely low level of the target peptide, T cells could be activated by NCI-H441 cells in the presence of the V2- U scDb, as evidenced by the secretion of IFNy and cytotoxicity (Fig. 42, A and B). The potency of V2-U was high, with ECso of 140 pM and 76 pM for IFNy secretion and cytotoxicity, respectively (Fig. 42, A and B). To rigorously assess the specificity of the V2- U scDb, the HLA-A3 allele was disrupted in NCI-H441 cells by CRISPR-based technologies and the knock-out (KO) was confirmed via flow cytometry (Fig. 54). KO of the HLA-A3 allele eliminated the ability of the V2-U scDb to elicit IFNy secretion or cytotoxicity by T cells upon exposure to the target NCI-H441 cells (Fig. 42, C and D). CRISPR-based technologies were then used to replace ("knock-in" [KI]) KRAS G12V with KRAS G13D in parental NCI-H441 cells containing their endogenous HLA-A3 allele (Fig. 55). G12V was replaced with G13D, rather than knocking out the G12V allele, to maintain viability of the cells, which require mutant KRAS genes. Two independent NCI-H441 clones with the KRAS G13D substitution were tested, and both substantially abrogated the ability of T cells to be activated by V2-U scDb in the presence of NCI-H441 cells (Fig. 42, C and D). Similarly, specific activation of T cells co-cultured with NCI-H441 cells was observed with the V2-U2 scDb, but as expected, the potency of the V2-U2 scDb was not as great as the V2-U scDb (Fig. 56). In all the experiments with endogenous levels of HLA-A3 and G12V alleles (Fig. 42, A and C, Fig. 56A), IFNy secretion from T cells was considerably lower than that in T cells activated by the transfected COS-7 cells (Fig. 41 A, Fig. 539C), consistent with greater numbers of the pHLA complexes on the transfected COS-7 cells. Even so, the scDbs were able to induce efficient NCI-H441 target cell lysis through T cell activation in a KRAS G12V and HLA-dependent fashion (Fig. 42, B and D, Fig. 56B). Moreover, other markers of T cell activation (TNFa, IL-2, granzyme B, and perforin) were released in a dose-dependent manner, demonstrating that the V2-U scDb was capable of inducing a poly-functional T cell response against cells expressing very low levels of antigen (Fig. 57).

To further assess the specificity of the V2-U scDb, a second cell line, NCI-H358, was used. This lung cancer cell line contains the HLA-A3 allele and a KRAS G12C mutation. Using CRISPR, the GUV mutation was introduced in the KRAS locus in three independent clones (Fig. 55). All three GUV clones were able to induce cytokine secretion from T cells in the presence of the V2-U scDb, while parental cells or a clone retaining the G12C allele were not (Fig. 42E). Cytotoxicity to the GUV clones was also much greater, especially at subnanomolar scDb concentrations, than to isogenic NCI-H358 cells without the GUV allele (Fig. 42F). Similarly, specific cytokine secretion and cytotoxicity was observed upon co culture of these cells with T cells in the presence of V2-U2 scDb (Fig. 58). All NCI-H358 variants expressed approximately the same level of HLA-A3 (Fig. 54).

IFNy secretion in co-cultures of T cells with several other HLA-A3+ cancer cell lines without RAS GUV mutations was assessed. These lines included A-427 (lung adenocarcinoma), COLO 741 (melanoma), Hs 578T (breast invasive ductal carcinoma),

Jurkat (acute T cell leukemia), SK-MES-1 (lung squamous cell carcinoma), and SW780 (bladder transitional cell carcinoma). CFP AC-1, the KRAS GUV and HLA-A3+ pancreatic adenocarcinoma cell line that presents an average of only ~3 copies of the GUV peptide per cell were also assessed (Table 10). Expression of ELLA- A3 in all these cell lines was confirmed via flow cytometry (Fig. 54). Neither V2-U nor V2-U2 scDb resulted in appreciable IFNy release from T cells co-cultured with the cancer cell lines without the KRAS GUV mutation (Fig. 59). V2-U scDb induced a low but significantly higher level of IFNy release with CFP AC- 1 cells (Fig. 59A); however, there was no statistically significant difference when V2-U2 was used (Fig. 59B). To study the ability of the L2-U scDb to induce T cell activation, co-culture with a panel of cell lines that differed in RAS mutation status and HLA-A1 expression were employed (Fig. 43 A, Fig. 60). Substantial dose-dependent IFNy secretion by T cells was only observed with the one cell line (HL-60) containing both HLA-A1 and RAS Q61L alleles (Fig. 43 A). As noted above, HL-60 presents an average of four Q61L pHLA complexes per cell. In a titration experiment, the L2-U scDb was able to induce IFNy release from T cells with an ECso of 60 pM (Fig. 43B) and HL-60 cell death with an ECso of 34 pM (Fig. 43C), despite the very low level of Q61L peptides per cell. Other markers of T cell activation were similarly released in a dose-dependent manner, showing that the L2-U scDb, like the V2-U scDb, was capable of inducing a poly-functional T cell response (Fig. 61). To assess the specificity of the L2-U scDb antigenic determinants in HL-60 cells, NRAS Q61H-KI, Q61R- KI, and HLA-A1-KO variants of these cells were generated (Fig. 55). These experiments confirmed that IFNy secretion (Fig. 43D) and target cell cytotoxicity (Fig. 43E) were dependent on the co-presence of NRAS Q61L and HLA-A1 genes in HL-60 cells.

Assessing potential cross reactivity to other putative HLA-A1 and HLA-A 3 -binding peptides To investigate whether the V2 scFv could bind to similar peptides derived from other proteins, a protein BLAST (BLASTp) search of the human RefSeq proteome was performed with the amino acid sequences of the G12V peptide or its G12WT, G12C, or G12D counterparts. Thirty-two proteins containing similar peptides were identified through this search. Of these, NetMHC v4.0 predicted that 17 peptides had strong or weak binding of HLA-A3 (Table 14). Each of these 17 peptides were synthesized and used to pulse T2A3 cells. While the majority of these peptides bind to HLA-A3, as assessed by GAP. A3 antibody staining, V2 phage only recognized the peptide IIVGAIGVGK (“Blast2”; SEQ ID NO:260), a peptide derived from the protein Rab-7b (Fig. 62A, Table 14).

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 14. RAS G12 BLAST peptides. BLASTp was used to query the human RefSeq proteome for peptides similar to the G12V, G12WT, G12C, and G12D [7-16] peptides. These similar peptides were extended with the flanking amino acids from the source protein to produce the full peptide sequence. These peptides were analyzed by NetMHCv4.0 for predicted weak and strong binders to HLA*A03:01. These predicted binders are listed as BLAST peptides.

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Rab-7b is a RAS-related protein expressed in monocytic cells and keratinocytes. To investigate whether this peptide represents an authentic alternative target and thus could cause off-target toxicity of the V2-U scDb, T cells first co-cultured with peripheral blood mononuclear cells (PBMC), monocytes, iDCs, and mature dendritic cells (mDC) prepared from an HLA-A3+ donor (Fig. 62, B and C). None of these cells were able to activate T cells significantly above background as assessed by IFNy secretion in the presence of the V2-U scDb (Fig. 62C). Importantly, these cells have the ability to activate T cells when pulsed with the G12V peptide (Fig. 62C). Next, V2-U scDb was tested against a skin derived cell line,

Hs 695T, which highly expresses Rab-7b (Fig. 62B). As these cells do not express the HLA- A3 allele, they were transfected with either HLA-A3 or HLA-A2 and co-cultured with T cells. Despite high levels of expression of endogenous Rab-7b, Hs 695T cells did not activate T cells in the presence of V2-U scDb. (Fig. 62D). As a positive control, the same experiment was performed after pulsing Hs 695 T cells with the G12V peptide, and under these circumstances, robust activation of T cells was achieved (Fig. 62D).

As a final assessment of the potential cross-reactivity with the Rab-7b peptide, plasmids encoding full length Rab-7b, KRAS WT, or KRAS G12V, in combination with HLA-A3, were transfected into COS-7. Cells overexpressing the mutant KRAS induced robust T cell activation in the presence of V2-U scDb (Fig. 62E), whereas COS-7 cells expressing control proteins, KRAS WT and Rab-7b, showed only marginal, non-dose dependent activation of the same T cells (Fig. 62, B and E). This experiment was repeated in HCT 116 cells transfected with the same genes and again found that T cells were activated by KRAS in a mutant-dependent fashion but no activation was observed in Rab-7b-expressing cells (Fig. 62, B and F).

To further assay for potential cross-reactivity of the V2-U and L2-U scDbs, their binding to libraries of positional scanning variant peptides was evaluated. The library of peptides was generated by systematically substituting each amino acid of the original peptides with the other 19 amino acids. This resulted in 190 variants for each of the G12V and Q61L peptides for the V2 and L2 scDbs. The variant peptides were pulsed on to T2A3 cells (for V2-U scDb) or SigM5 cells (for L2-U scDb) and co-cultured with T cells. Recognition of the variant peptides was evaluated through IFNy release (Fig. 44, A and B, Fig. 63, A and B). For both scDbs, amino acid positions in the C-terminal half of the peptides, where the mutant residue resided, demonstrated greater specificity. Most changes of amino acids at these positions abolished recognition by the scDbs. On the other hand, amino acids at the N-terminus of the peptides could in many cases be substituted without substantially changing interaction with the scDbs. These recognition patterns are also illustrated as Seq2Logo graphs (Fig. 44, C and D). Next, using a 20% cognate peptide reactivity as a cutoff for permissive amino acids at each position, decamer binding motifs were generated for each candidate target peptide (Materials and Methods). A search of these motifs in the UniProtKB human protein database using ScanProsite yielded 162 peptides (including Rab-7b IIVGAIGVGK (SEQ ID NO:260)) that could potentially bind to V2 and 232 peptides that could potentially bind to L2 (Tables 15 and 16). Comparing these peptides to an extensive database of peptides actually presented by HLA as assessed by mass spectrometry, it was found that none of the 162 peptides and only one of the 232 peptides were known to be presented by the cognate HLA. However, when this one peptide (DTELQGMNEY (SEQ ID NO:310), from chromodomain-helicase-DNA-binding protein 4 [CHD4]) was pulsed on SigM5 cells, which was then co-cultured with T cells and L2-U scDb, it did not elicit IFNy release (Fig. 63C), demonstrating that L2-U scDb did not bind the CHD4 peptide. Table 15. Human peptides matching the V2 binding motif. Using a 20% cognate peptide reactivity as a cutoff for permissive amino acids at each position based on the results of the peptide library analysis, a decamer binding motif for V2 was generated. These motifs were searched in the UniProtKB human protein database using ScanProsite to identify peptides matching the motif.

Table 16. Human peptides matching the L2 binding motif. Using a 20% cognate peptide reactivity as a cutoff for permissive amino acids at each position based on the results of the peptide library analysis, a decamer binding motif for L2 was generated. These motifs were searched in the UniProtKB human protein database using ScanProsite to identify peptides matching the motif.

Evaluation of scDbs in mouse models

To determine whether the L2-U scDb could control tumor growth in vivo , HL-60 leukemia cells ( NRAS Q61L/WT) and human T cells were intravenously injected into NSG mice to establish widespread leukemic infiltrates. As controls, HL-60 cells engineered to harbor Q61H instead of Q61L/WT alleles were used in a separate group of mice. Bioluminescence established tumor uptake, and mice were randomized to receive the L2-U scDb or a control scDb through intraperitoneal 14-day infusion pumps. The L2-U scDb slowed the growth of the Q61L leukemic cells (Fig. 64C), but not the HL-60 Q61H cells (Fig. 64D). Though the tumor growth retardation was significant (p<0.05, multiple t-tests with Bonferroni-Dunn correction), the effect size was modest. Therapeutic effects were not assessed at later time points because of the lifetime of the pumps. No substantial changes in body weight were noted in any of the V2-U or L2-U scDb treated mice (Fig. 65).

Together these results demonstrate that highly specific bispecific antibodies can be generated against pHLA complexes resulting from common mutations occurring in cancer cells. The format and configuration of the bispecific antibodies developed here are highly specific and sensitive scDbs against protein products containing mutations occurring in cancer cells.

Materials and Methods

Study Design

The objective of this study was to generate therapeutic agents that target common mutations in RAS genes. This was accomplished by using phage display to identify scFvs specific to MANAs that had been confirmed to be presented via mass spectrometry. These scFvs were grafted into an optimized bispecific antibody format, the scDb. The scDbs were then shown to mediate MANA-specific T cell activation and target cell cytotoxicity in overexpression and endogenous-level expression model systems. All data presented were representative of data collected during this study. All experiments were performed in triplicate with three technical replicates unless otherwise noted. All experiments were performed in a way to minimize confounding variables, such as plate layout effects.

Plasmids

Plasmids encoding KRAS (isoform B), HRAS (isoform 1), and NR.AS variants (WT and mutant) and HLA class I alleles A*01:01 and A*03:01 were synthesized and cloned into pcDNA3.1 by GeneArt (Thermo Fisher Scientific, Waltham, MA) or synthesized into gBlocks (IDT, Coralville, Iowa) and assembled into pcDNA3.4 using NEBuilder® HiFi DNA Assembly Cloning Kit (NEB, Ipswich, MA).

Cell Lines and Primary Cells

All cells were grown at 37°C under 5% CO2. HEK293F cells (Thermo Fisher Scientific) were cultured in FreeStyle Expression media. T2A3 cells (a kind gift from Eric Lutz and Elizabeth Jaffee, JHU) were cultured in RPMI-1640 (ATCC, Manassas, VA) with 10% HyClone FBS (GE Healthcare, Chicago, IL), 1% Penicillin-Streptomycin (Thermo Fisher Scientific), 500 pg/mL Geneticin (Thermo Fisher Scientific), and IX Non-Essential Amino Acids (Thermo Fisher Scientific). COS-7, NCI-H441, CFP AC-1, NCI-H358, and HCT 116 cells (all from ATCC, Manassas, VA) were cultured in McCoy’s 5 A (Modified) (Thermo Fisher Scientific) with 10% HyClone FBS and 1% Penicillin-Streptomycin. Jurkat (Clone E6-1, ATCC), COLO 741 (Sigma- Aldrich, St. Louis, MO), and SW780 (ATCC) cells were cultured in RPMI-1640 with 10% HyClone FBS and 1% Penicillin-Streptomycin. KMS-21-BM (JCRB Cell Bank, Osaka, Japan) cells were cultured in RPMI-1640 with 20% HyClone FBS and 1% Penicillin-Streptomycin. A-427, Hep G2, Hs 695T, SK-MES-1 (all from ATCC) cells were cultured in Eagle's Minimum Essential Medium (ATCC) with 10% HyClone FBS and 1% Penicillin-Streptomycin. SigM5 (DSMZ) and HL-60 (ATCC) cells were cultured in Iscove's Modified Dulbecco's Medium (ATCC) with 20% HyClone FBS and 1% Penicillin-Streptomycin. RD (ATCC) cells were cultured in Dulbecco's Modified Eagle's Medium (ATCC) with 10% HyClone FBS and 1% Penicillin-Streptomycin. Hs 578T (ATCC) cells were cultured in Dulbecco's Modified Eagle's Medium (ATCC) with 10% HyClone FBS, 1% Penicillin-Streptomycin, and 0.01 mg/ml bovine insulin (Sigma-Aldrich).

Peripheral blood cells were obtained from healthy volunteer donors or purchased as leukapheresis samples (Stem Cell Technologies, Vancouver, BC). PMBCs were purified by density gradient centrifugation with Ficoll Paque Plus (GE Healthcare). T cells were expanded from PBMCs with addition of the anti-human CD3 antibody (clone OKT3, BioLegend, San Diego, CA) at 15 ng/mL, or with Human T-Activator CD3/CD28 Dynabeads (Thermo Fisher Scientific) at a 1:5 beadxell ratio for three days, after which beads were removed with a magnet and the medium was changed. T cells were cultured in RPMI-1640 with 10% HyClone FBS, 1% Penicillin-Streptomycin, 100 IU/mL recombinant human IL-2 (aldesleukin, Prometheus Therapeutics and Diagnostics, San Diego, CA), and 5 ng/mL recombinant human IL-7 (BioLegend). The culture medium was changed every 3-4 days and cells were maintained at ~1 million cells/mL.

To generate dendritic cells, monocytes were negatively isolated from PBMCs using microbeads (Miltenyi) and cultured in Mo-DC differentiation media (Miltenyi) for 5 days to induce differentiation into immature dendritic cells. To generate mature dendritic cells, immature dendritic cells were cultured with 0.5 mg/mL CD40 ligand oligomer (Enzo) in Mo-DC differentiation media for 2 more days.

Phage Display Library Construction

All cloning was modeled in SnapGene (GSL Biotech LLC, San Diego, CA). Oligonucleotides were synthesized by GeneArt (Thermo Fisher Scientific) using trinucleotide mutagenesis (TRIM) technology. The oligonucleotides were incorporated into the pADL-lOb phagemid (Antibody Design Labs, San Diego, CA) (Fig. 66).

Ten ng of the ligation product was mixed on ice with 10 pL of electrocompetent SS320 cells (Lucigen, Middleton, WI) and 14 pL of double-distilled water. This mixture was electroporated (200 ohms, 25 microFarads, 1.8 kV) using a Gene Pulser electroporation system (Bio-Rad, Hercules, CA) and allowed to recover in Recovery Media (Lucigen) for 45 minutes at 37°C. Cells transformed with 60 ng of ligation product were pooled and plated on a 24-cm x 24-cm plate containing 2xYT medium supplemented with carbenicillin (100 pg/mL) and 2% glucose. Cells were grown at 37°C for 6 hours and placed at 4°C overnight. To determine the transformation efficiency for each series of electroporations, aliquots were titered by serial dilution. Cells grown on plates were scraped into 850 mL of 2xYT medium with carbenicillin (100 pg/mL) plus 2% glucose for a final Oϋόoo of 5-15. Two mL of the 850 mL culture were taken and diluted ~1 :200 to reach a final Oϋόoo of 0.05-0.07. To the remaining culture, 150 mL of sterile glycerol was added before snap freezing to produce glycerol stocks. The diluted bacteria were grown to an Oϋόoo of 0.3-0.5, transduced with M13K07 helper phage at an MOI of 4 (Antibody Design Labs) and shaken at 37°C for 1 hour. The culture was centrifuged and the cells were re-suspended in 2xYT medium with carbenicllin (100 pg/mL), kanamycin (50 pg/mL), and IPTG (50 mM, Thermo Fisher Scientific) and grown overnight at 30°C for phage production. The following morning, the bacterial culture was aliquoted into 50 mL Falcon tubes and pelleted twice at high speed to obtain clarified supernatant. The phage-laden supernatant was precipitated on ice for 40 minutes with a 20% PEG-8000 / 2.5 M NaCl solution at a 1 :4 ratio of PEG/NaCl: supernatant. After precipitation, phage were centrifuged at 12,000 g for 40 minutes and re-suspended in lx TBS (25 mM Tris-HCl, 150 mM NaCl, pH 7.5) containing 2 mM EDTA. Phage from multiple tubes were pooled and re-precipitated to achieve a higher concentration. Final phage were re-suspended in lx TBS, 2 mM EDTA, and lx Complete Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO). The total number of transformants obtained was determined to be 3.6 x 10¹⁰. The library was aliquoted and stored in 15% glycerol at -80°C and in 50% glycerol at -20°C.

DNA from the library was amplified using the following primers (Forward: CGACGTAAAACGACGGCCAGTNNNNNNNNNNNNNNCGTGCAGAGGATACAGC AGTG (SEQ ID NO: 681), Reverse:

CACACAGGAAACAGCTATGACCATGCTAACGGTAACCAGGGTGCCCTG (SEQ ID NO:682)) which flank the CDR-H3 region. (All oligonucleotide sequences listed in this manuscript begin with the most 5' nucleotide.) The sequences at the 5'-ends of these primers incorporated molecular barcodes to facilitate unambiguous enumeration of distinct phage sequences as well as universal primer sites. The protocols for PCR-amplification and sequencing were described elsewhere (Kinde et ak, Proc Natl Acad Sci USA 108:9530-9535 (2011)). Sequences were processed and translated using a custom SQL database and both the nucleotide sequences and amino acid translations were assessed. Peptides and pHLAs

Peptides (Table 11) were synthesized at a purity of >90% by Peptide 2.0 (Chantilly, VA) or ELIM Biopharm (Hayward, CA), with the exception of the crude peptides that were used for the positional scanning library. Peptides were re-suspended in DMF at 10 mg/mL and stored at -80°C. pHLAs were synthesized by refolding recombinant HLA-A*01 :01 (HLA-A1) or HLA-A*03:01 (HLA-A3) with peptide and beta-2 microglobulin, purified by gel-filtration, and biotinylated (Fred Hutchinson Immune Monitoring Lab, Seattle, WA; or Baylor MHC Tetramer Production Lab, Houston, TX). These pHLA were confirmed to be folded prior to selection via ELISA using the W6/32 antibody (BioLegend, San Diego, CA), which recognizes only folded HLA. Blast peptides (Table 14) were synthesized as described above, re-suspended in DMF at 10 mg/mL and stored at -80°C. Cognate peptide reactivity search of the UniProtKB human protein database using ScanProsite was performed using binding motifs with a 20% parental peptide IFNy value as a cutoff. The V2 motif was {FWDY}-[ILMVTC]-{RE}-{ILV}-x-[ILV]-[GNST]-[VP]-[AG]-[HKY] (SEQ ID NO:683). The L2 motif was x-{PWRHDEY}-[APRDEQSC]-{DE}-[AMFPGHDNQSTYC]-[AG]- [ILM] - [ AIMGRDENQ] - [DE] - [AY] (SEQ ID NO:684).

Selection of mutant pHLA-specific phage clones scFv-bearing phage specific to the RAS G12V[7-16] “G12V” pHLA-A3 and RAS Q61H, Q61L, Q61R, referred to collectively as “Q61X”, in pHLA-Al were identified using methods similar to those described elsewhere (Skora et al., Proc Natl Acad Sci USA 112:9967-9972 (2015); and Miller et al., J Biol Chem 294: 19322-19334 (2019)). The phage display library was regrown within a week of starting the selection process. A colony of phage-competent SS320 cells (Lucigen, Middleton, WI) was inoculated in 2xYT medium (Sigma-Aldrich, St. Louis, MO) supplemented with tetracycline (20 pg/mL) and cultured at 37°C overnight, then grown to 2 L of mid-log phase (OD₆oo of 0.3-0.5) bacteria. Bacteria were transduced with the phage library at an MOI of 0.5 and M13K07 helper phage (Antibody Design Labs, San Diego, CA) at an MOI of 4 along with the addition of 2% (W/V) glucose (Sigma-Aldrich, St. Louis, MO) and shaken, not stirred, for 1 hour at 37°C. The cells were pelleted and re-suspended in 2xYT medium with carbenicillin (100 pg/mL), kanamycin (50 pg/mL), and 50 pM IPTG and subsequently shaken and grown overnight at 30°C for phage production. The following morning, the bacterial culture was aliquoted into 50 mL Falcon tubes and pelleted twice by centrifugation at 12,000 x g to obtain clarified supernatant. The phage-laden supernatant was precipitated on ice for 40 minutes with a 20% PEG-8000 / 2.5 M NaCl solution at a 1 :4 ratio of PEG/NaCl: supernatant. After precipitation, phage were pelleted by centrifugation at 12,000 x g for 40 minutes and re-suspended in 1 mL of lx TBS with 2 mM EDTA, 0.1% sodium azide, and lx Complete Protease Inhibitor Cocktail (Sigma-Aldrich, St. Louis, MO).

For the G12V pHLA-A3 target MANA, the selection scheme involved an enrichment phase (one round), a competition phase (up to three rounds), and a final selection phase (two rounds). The biotinylated pHLA were incubated with 25 pL of M-280 streptavidin Dynabeads (Invitrogen, Thermo Fisher Scientific) or 100 pL of streptavi din-coated agarose beads (Novagen EMD Millipore, Burlington, MA) per 1 pg of pHLA in blocking buffer (phosphate-buffered saline or PBS, 0.5% BSA, 0.1% sodium azide) for 1 hour at room temperature (RT). After the initial incubation, the complexes were washed and re-suspended in 100 pL of blocking buffer. During the enrichment phase (Round 1), approximately 4 x 10¹² phage, representing ~ 100-fold coverage of the library, were negatively selected overnight at 4°C against 500 pL unconjugated washed Dynabeads, 500 pg free streptavidin protein (RayBiotech, Norcross, GA), and 3 pg heat-denatured, allele-matched HLA conjugated to Dynabeads. This step was designed to remove phage recognizing Dynabeads, streptavidin or denatured HLA- A3. After negative selection, beads were isolated with a DynaMag-2 magnet (Life Technologies, Carlsbad, CA) and the supernatant containing unbound phage was used for positive selection by incubation with 0.5 pg G12V pHLA-A3 conjugated to Dynabeads for 1 hour at RT. The beads were washed 10 times with lx TBS with 0.5% Tween-20 using the DynaMag-2 magnet, and the phage were eluted from the beads by re-suspension in 1 mL of 0.2 M glycine, pH 2.2. After a 10-minute incubation, the solution was pH-neutralized by the addition of 150 pL of 1 M Tris, pH 9.0. This neutralized solution was used to transduce 10 mL cultures of mid-log-phase SS320s to which M13K07 helper phage (MOI of 4) and 2% glucose was added. Bacteria were then incubated as described above and the phage precipitated the next morning with PEG/NaCl.

During the competitive phase (Rounds 2-4), the amount of input phage used in each round was decreased to 5%, 1%, and 0.1% of the total precipitated phage from the previous round, respectively. These phage were subjected to negative selection against 1 pg heat- denatured HLA-A3, 1 pg total of unrelated pHLA-A3, and 20 pg free streptavidin for 1 hour at RT. After negative selection, beads were isolated with a DynaMag-2 magnet and the unbound phage were used for positive selection. This was accomplished by simultaneously co-incubating phage with 0.5 pg G12V pHLA-A3 conjugated to the magnetic Dynabeads and corresponding 1 pg G12WT pHLA-A3 conjugated to streptavidin-coated agarose beads as competitor. Prior to elution, beads were washed 10 times in 1 mL lx TBS with 0.5% Tween- 20. Phage were eluted from magnetic Dynabeads and used to transduce mid-log phase SS320 cells as described above.

During the final selection phase, phage resulting from rounds 2, 3, and 4 were separately subjected to additional, more stringent rounds of selection. 0.1% of the precipitated phage from these rounds underwent two negative selections against 0.5 pg G12WT pHLA-A3, followed by a positive selection against 0.5 pg of G12V pHLA-A3. The beads were washed 10 times in 1 mL lx TBS with 0.5% Tween-20, and phage were eluted and used to transduce mid-log phase SS320 cells as described above. The final selection steps described above was repeated a second time, thus the phage underwent a total of four to six total rounds of negative/positive selection. scFv-bearing phage specific to RAS Q61X-HLA-A1 MANA targets were selected as described above with the following differences. The Q61X pHLA-Al panning scheme involved one round of enrichment and four rounds of more stringent selection. In the enrichment round, ~2.6 x 10¹³ phage, representing ~720-fold coverage of the library, were negatively selected against 1 mL unconjugated washed Dynabeads and 1 mg free streptavidin protein. This was followed by positive selection of the unbound phage using 2 pg of the mutant Q61X pHLA-Al. For the four subsequent selection rounds, 10%, 1%, 0.1%, and 0.02% of the phage from the previous round were used as input for panning, respectively. In each of these rounds, phage were negatively selected using 2 pg Q61WT pHLA-Al, 2 pg unrelated pHLA-Al, and 5 x 10⁸ cells from HLA-A1+ cell lines lacking the RAS mutation of interest. Unbound phage were used for positive selection with 1 pg, 0.5 pg, 0.5 pg, and 0.25 pg Q61X pHLA-Al in the four sequential rounds, respectively.

To obtain monoclonal phage, individual colonies of SS320 cells transduced with a limiting dilution of phage were inoculated into 200 pL of 2xYT medium containing 100 pg/mL carbenicillin and 2% glucose and grown for 3 hours at 37°C. The cells were then transduced with 1.6 x 10⁷ M13K07 helper phage and incubated for 1 hour at 37°C with shaking. The cells were pelleted, re-suspended in 300 pL of 2xYT medium containing carbenicillin (100 pg/mL), kanamycin (50 pg/mL), and 50 pM IPTG, and grown overnight at 30°C for phage production. Cells were pelleted and the phage-laden supernatant was used for downstream analysis.

PCR and Sanger Sequencing

Monoclonal phage DNA was PCR amplified using 1 pL of monoclonal phage supernatant in a reaction with primers flanking the CDRs (Forward: GGCCATGGCAGATATTCAGA (SEQ ID NO: 198), Reverse: CCGGGCCTTTATCATCATC (SEQ ID NO: 199)) and Q5 Hot Start High-Fidelity 2X Master Mix (New England BioLabs). The PCR product was Sanger-sequenced by Genewiz (South Plainfield, NJ). Sequences flanking the CDRs were trimmed using DNA Baser Sequence Assembler v4 (Arges, Romania) and the sequences spanning the CDRs were clustered using the CD-HIT Suite. Colonies containing unique phage clones were selected for further assays.

ELISAs

Streptavidin-coated, 96-well plates (R&D Systems, Minneapolis, MN) were coated with 50 ng of biotinylated pHLA-A3 or pHLA-Al (unless otherwise specified) or 25 ng of biotinylated recombinant heterodimeric CD3e/5 (Abeam, Cambridge, MA) in 50 pL of blocking buffer (PBS with 0.5% BSA, 2 mM EDTA, and 0.1% sodium azide) at 4°C overnight. Plates were washed with lx TBST (lx TBS + 0.05% Tween-20) using a BioTek 405 TS plate washer (BioTek, Winooski, VT).

The phage clones resulting from the RAS G12V pHLA-A3 panning were characterized via monoclonal ELISA, where individual monoclonal phage clones were separately interrogated for their binding to G12V pHLA-A3 and G12WT pHLA-A3 via ELISA. 50 pL of phage supernatant was added to washed streptavidin ELISA plates coated with either G12V or G12WT pHLA-A3. Plates were incubated for 2 hours at RT and then washed 6 times. The bound phage were then incubated with 50 pL of rabbit anti-fd/M13 bacteriophage antibody (Novus Biologicals, Abingdon, UK) diluted 1:3000 in lx TBST for 1 hour at RT, followed by washing and incubation with 50 pL of goat anti-rabbit HRP (Thermo Fisher Scientific) diluted 1 : 10,000 in lx TBST for 1 hour at RT. After washing, 50 pL of 3,3',5,5'-tetramethylbenzidine (TMB) substrate (BioLegend) was added to each well and the reaction was quenched with 1 N sulfuric acid (Fisher Scientific, Thermo Fisher Scientific). Absorbance at 450 nm was measured with a Synergy HI Multi-Mode Reader (BioTek).

ELISA with purified scFvs, scDbs, and other bispecific antibody formats was performed essentially as above, with serial dilutions of the recombinant protein of interest incubated for 1 hour at RT, followed by incubation with 1 pg/mL recombinant protein L (Pierce, Thermo Fisher) for 1 hour at RT, followed by incubation with anti-protein L HRP (Abeam). Plates were washed, exposed, and read as described above. pHLA titration ELISAs assessing the binding of the scFvs (at a fixed concentration) to the mutant and WT pHLA (at varying concentrations) were performed by AxioMx Inc (Abeam).

Sandwich ELISAs were performed by coating biotinylated pHLA- A3 on a streptavidin plate and incubating with scDbs as described above, followed by incubation with recombinant heterodimeric CD3e/5 protein containing a human Fc domain at 1 pg/mL for 1 hour at RT, followed by detection with anti-human Fc HRP (Abeam) at 1:10,000 for 1 hour at RT. Plates were washed, exposed, and read as described above.

Flow Cytometry

For peptide pulsing of cells, cells were washed once with PBS and once with RPMI- 1640 containing 1% Penicillin- Streptomycin without serum. The cells were then incubated at 5 x 10⁵ - 1 x 10⁶ cells per mL in serum-free RPMI-1640 containing 50 pg/mL or specified concentration of peptide and 10 pg/mL human beta-2 microglobulin (ProSpec, East Brunswick, NJ) for 4 hours at 37°C. Prior to staining, cells were spun and re-suspended in cold stain buffer (PBS containing 0.5% BSA, 2 mM EDTA, and 0.1% sodium azide).

The phage clones resulting from the RAS Q61X pHLA-Al selection were characterized via flow cytometry, where individual monoclonal phage clones were separately interrogated for their binding to mutant Q61X and Q61WT peptide-pulsed SigM5 cells. Monoclonal phage were grown and sequenced as described above. Phage supernatant from representative wells of each unique clone was selected for flow cytometry analysis. In each well of a deep 96-well 2 mL plate, 50 pL of monoclonal phage was incubated with 2.5 x 10⁵ peptide-pulsed SigM5 cells in 50 pL of stain buffer. Plates were incubated on ice for 1 hour, followed by washing with 1 mL of stain buffer. Cells were then stained with 1 pg of rabbit anti -Ml 3 antibody (Novus Biologicals), washed, stained with anti-rabbit-PE (BioLegend), incubated with an additional 100 pL of LIVE/DEAD Fixable Near-IR dye diluted 1 : 1000 in PBS for 10 minutes at RT in the dark, followed by washing in stain buffer before analysis. Stained cells were analyzed using an Intellicyt iQue3 flow cytometer (Sartorius, Gottingen, Germany).

Peptide-pulsed T2A3 phage staining assays were performed by incubating 5 x 10⁵ - 1 x 10⁶ cells with 1 x 10¹⁰ phage in 100 pL stain buffer on ice for 1 hour, followed by one wash in stain buffer. Cells were then stained with 1 pg of rabbit anti -Ml 3 antibody (Novus Biologicals), washed, stained with anti-rabbit-PE (BioLegend), incubated with an additional 500 pL of LIVE/DEAD Fixable Near-IR dye diluted 1:1000 in PBS for 10 minutes at RT in the dark, and washed in stain buffer before analysis. V2 scFv staining was performed using 0.33 pg of V2 scFv premixed with 1 pg anti-FLAG-PE antibody (BioLegend) overnight at 4°C followed by incubation with peptide-pulsed T2A3 cells as described above. Anti-HLA-A3 staining was performed by staining 5 x 10⁵ cells with 0.125 pg clone GAP.A3-PE (eBioscience, Thermo Fisher Scientific) or mouse isotype IgG2a-PE (BioLegend). Anti-HLA-Al staining was performed by staining 5 x 10⁵ cells with 0.5 pg anti-HLA-Al/Al 1/A26 antibody clone 8.L.101 (Abeam) or mouse isotype IgM (Thermo Fisher) followed by 0.25 pg anti-mouse-PE (BioLegend). Stained cells were analyzed using an LSRII flow cytometer (Becton Dickinson, Mansfield, MA).

Quantification of cell surface-bound G12V peptide was performed using two commercial kits: PE Quantibrite Beads (BD, Franklin Lakes, NJ) and QIFIKIT (Agilient, Santa Clara, CA). For Quantibrite-based quantification, peptide-pulsed T2A3s were stained with 0.5 pg V2 scFv preconjugated to 1.5 pg clone L5 anti-FLAG-PE (BioLegend), followed by flow cytometry and quantification according to the manufacturer’s instructions. For QIFIKIT-based quantification, peptide-pulsed T2A3s were stained with 0.5 pg V2 scFv preconjugated to 1.5 pg clone M2 anti -FLAG (Sigma Aldrich), followed by staining with anti-mouse-PE (BioLegend), flow cytometry and quantification according to the manufacturer’s instructions. Recombinant scFv and bispecific antibody production

Recombinant scFv proteins were produced and purified by AxioMx Inc (Abeam). In brief, the V2, HI, L2, and R6 scFv sequences were subcloned into a vector containing a periplasmic localization sequence, and C-terminal Flag and His tags. ScFvs were expressed in E. coli and purified via nickel chromatography.

Bispecific antibodies were produced after subcloning gBlocks (IDT, Coralville, Iowa) encoding each of the variants with an IL-2 signal sequence and C-terminal 6xHIS tag into the pcDNA3.4 vector (Thermo Fisher Scientific). Bispecific antibodies were produced by the Eukaryotic Tissue Culture Core Facility of Johns Hopkins University. In brief, 1 mg of plasmid was transfected using polyethylenimine (PEI) into HEK293F cells which were then grown as suspension culture at a density of 2 x 10⁶ cells/mL in FreeStyle 293 expression media (Thermo Fisher Scientific) at 37°C, 170 rpm, and 5% CO2. Protein was expressed for 5 days, after which cells were harvested by centrifugation, and the resulting supernatant filtered using a 0.22 pm filter. To each 1 L of supernatant, 2 mL of Ni-NTA His-Bind (Millipore Sigma) resin slurry was added and allowed to incubate at 4°C overnight on an orbital shaker. The supernatant containing the slurry was passed through a centrifuge column (Pierce, Thermo Fisher Scientific), whereby the slurry was washed with 20 mM imidazole in PBS, and eluted in 50 mM, 100 mM, and 250 mM imidazole fractions. Protein fractions were run on mini-PROTEAN TGX stain-free gels (Bio-Rad, Hercules, California) and appropriate fractions were combined for desalting into PBS, pH 7.4 or 20 mM Tris, 150 mM NaCl, pH 9 using 7k MWCO Zeba Spin desalting columns (Thermo Fisher Scientific). Proteins were quantified via stain free gel and BCA protein assay (Pierce, Thermo Fisher Scientific). For scDb western blots, protein was transferred from the stain-free gel to a Trans-Blot Turbo Mini 0.2 pm PVDF membrane (Bio-Rad) using the Trans-Blot Turbo Transfer System (Bio-Rad). Membrane was blocked with blocking buffer (5% Bio-Rad Blotting-Grade Blocker in lx TBST) on an orbital shaker (VWR) for 1 hour at RT, followed by incubation with anti-6x His tag antibody clone ab9108 at 1 : 1000 in blocking buffer at 4°C overnight, washing in lx TBST, and incubation with anti-rabbit-HRP (Thermo Fisher Scientific) at 1 : 10,000 in blocking buffer for 1 hour at RT. Membrane was washed in lx TBST followed by ddH20, then imaged using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific) on a ChemiDoc XRS+ imager (Bio- Rad). Proteins were stored at 4°C for short term storage or snap frozen with the addition of 7% glycerol and stored at -80°C for long term storage. Alternatively, the V2-U scDb and L2- U scDb proteins were produced by GeneArt in Expi293s, purified with a HisTrap column followed by size exclusion chromatography using a HiLoad Superdex 20026/600 column.

SPR

RAS G12V pHLA-A3, G12WT pHLA-A3, and V2 scFv binding experiments were performed at 25°C using a Biacore T200 SPR instrument (GE Healthcare, Chicago, IL). Approximately 130-140 response units (RU) of biotinylated pHLA-A3 were captured in flow cells (Fc) 2 and 4, respectively, using a streptavidin chip. Single-cycle kinetics were performed by injecting increasing concentrations (1.56, 6.25, 25, 100, 400 nM) of purified V2 scFv flowed over Fc 1-4. Binding responses for kinetic analysis were both blank and reference subtracted. Both binding curves were fit with a 1 : 1 binding model using Biacore Insight evaluation software. Measurements for L2-U scDb were done similarly, using biotinylated Q61L pHLA-Al and Q61WT pHLA-Al. pHLA Immunoprecipitation and Mass Spectrometry pHLA immunoprecipitation and mass spectrometry were performed as described elsewhere (Wang et ak, Cancer Immunol Res 7:1748-1754 (2019)). Briefly, COS-7 cells seeded into 24.5 x 24.5 cm² plates were transfected at 95% confluency using Lipofectamine 3000 Reagent (Thermo Fisher Scientific). For each plate, 125 pg of plasmids (50 pg of HLA plasmid and 75 pg of mutant or WT protein plasmid) were mixed with 200 pL of Lipofectamine P3000 in 6 mL of Opti-MEM (Thermo Fisher Scientific). In a separate tube, 200 pL of Lipofectamine 3000 Reagent was mixed with 6 mL of Opti-MEM. The contents of the two tubes were mixed and allowed to complex for 10 minutes. Medium bathing cells were removed and 50 mL of fresh complete medium was added followed by the Lipofectamine-DNA mixture. Cells were harvested 48 hours post-transfection. The transfection efficiency of COS-7 was >90% as assessed by GFP+ cell fraction on flow cytometry (BD LSRII).

Cells (transfected or untransfected) were grown to near confluency in 24.5 x 24.5 cm² plates. Cultured cells were washed with PBS two times, followed by another wash with PBS pre-chilled at 4°C containing lx protease inhibitor. Cells were scraped and collected in a 500-mL centrifuge bottle. The bottle was centrifuged at 1,000 g for 5 minutes and the supernatant discarded. Cell pellets were snap frozen in liquid nitrogen and stored at -80°C for future experiments.

Neoantigen-expressing cells were processed as described elsewhere (Wang et al., Cancer Immunol Res 7: 1748-1754 (2019)). In brief, a total of 500 million cells were lysed and pHLA complexes were immunoprecipitated using Protein G Dynal Magnetic Beads (Thermo Fisher Scientific) pre-conjugated with anti-human HLA-A, B, C antibody clone W6/32 (Bio-X-Cell). After elution, dissociation, and filtration, peptides were lyophilized before further analysis. HPLC fractionation and a Dual-Reduction procedure were then performed (Wang et al., Cancer Immunol Res 7: 1748-1754 (2019)). Controls for detection of the RAS G12V [7-16] and [8-16] peptides were established using AQUA™ heavy isotope labeled peptides of the same sequence (Sigma-Aldrich). These AQUA peptides were added to the cell lysates in every experiment. Transition parameters were manually examined and curated to exclude ions with excessive noise due to co-elution with impurities. Absolute copy numbers of peptides presented on the cell surface were calculated based on the MANA- SRM quantification using the AQUA™ heavy isotope labeled peptides and the recovery ratios of the pipeline, as described elsewhere (Wang et al., Cancer Immunol Res 7:1748-1754 (2019)).

CRISPR on Cell Lines

The Alt-R CRISPR system (Integrated DNA Technologies, IDT) was used to modify the HLA alleles, the KRAS mutation status of the NCI-H358 and NCI-H441 cell lines, and the NRAS mutation status of the HL-60 cell line. Alt-R® CRISPR Cas9 crRNAs (IDT) and Alt- R® CRISPR-Cas9 tracrRNA (IDT) were re-suspended at 100 mM with Nuclease-Free Duplex Buffer (IDT). The crRNAs and tracrRNA were mixed at a 1 : 1 molar ratio and denatured for 5 minutes at 95°C, followed by slow cooling to room temperature to duplex prior to mixing with Cas9 Nuclease (IDT) at a 1.2:1 molar ratio for 15 minutes. To knock out the HLA-A3 allele in NCI-H441 cells, 40 pmoles of the Cas9 ribonucleoprotein (RNP) containing tracrRNA/HLA-A crRNA (GCTGCGACGTGGGGTCGGAC; SEQ ID NO:685) duplex were mixed with 2 x 10⁵ NCI-H441 cells in 20 pL of OptiMEM. This mixture was loaded into a 0.1 cm cuvette (Bio Rad) and electroporated at 150 V for 10 ms using an ECM 2001 (BTX). HLA-A1 was similarly knocked out in HL-60 cells by mixing Cas9 RNP containing tracrRNA/HL A- A 1 crRNA (CAGACTGACCGAGCGAACCTG; SEQ ID NO:686) duplex with 5 x 10⁵ HL-60 cells, and electroporated at 120V for 16 ms. Cells were immediately transferred to complete growth medium and cultured for 10 days.

To change ("KI") the KRAS mutation status of NCI-H358 from G12C/WT to G12V and NCI-H441 from G12V/WT to G13D/WT, Cas9 RNPs were co-el ectroporated with single-strand DNA homology directed repair templates. The G12V repair template was ATTAGCTGTATCGTCAAGGCACTCTTGCCTACGCCAACGGCGCCGACAACGACG AGTTTATATTCAGTCATTTTCAGCAGGCCTTATAA (SEQ ID NO:687) for KRAS- G12V crRNA (AATGACTGAATATAAACTTG; SEQ ID NO:688) . The G13D repair template was

AACAAGATTTACCTCTATTGTTGGATCATATTCGTCCACAAAATGATTCTGAATT AGCTGTATCGTCAAGGCACTCTTGCCTACGTCACCAGCTCCAACTACCACAAGTT TATATTCAGTCATTTTC (SEQ ID NO:689) for X&EVG13D-1 crRNA (CTTGTGGTAGTTGGAGCTGT; SEQ ID NO:690). Both repair templates were obtained as Ultramer® Oligos from IDT. To improve rates of homology directed repair, NCI-H441 cells were pre-treated with 200 ng/mL of nocodazole (Sigma Aldrich) and 1 mM of SCR7 pyrazine (Sigma Aldrich) for 17 hours prior to electroporation. The electroporation mixture contained 40 pmoles of Cas9 RNP, 20 pmoles of repair template, and 2 x 10⁵ cells in 20 pL of OptiMEM. NCI-H441 cells were electroporated at 150 V for 10 ms, while NCI-H358 cells were electroporated at 120 V for 16 ms. Both cell types were transferred to complete growth media containing 1 pM SCR7 for 72 hours following electroporation. Cells were grown in culture for 5 to 10 more days before use.

The NRAS mutation status in HL-60 was modified similarly. 5 x 10⁵ HL-60 cells were co-electroporated at 120 V for 16 ms with Cas9 RNPs containing tracrRNA/A7 T5' crRNA (CCTCATGTATTGGTCTCTCATGG; SEQ ID NO: 691) duplex and either the repair template for Q61H

(AAACCTGTTTGTTGGACATACTGGATACAGCTGGACATGAGGAATATTCTGCAA T GAGAGACC AAT AC AT GAGGAC AGGCGAAGGCTTCCT; SEQ ID NO:692) or Q61R ( AAACCTGTTT GTT GGAC AT ACTGGAT AC AGCTGGAAGAGAGGA AT ATTCTGC AA T GAGAGACC AAT AC AT GAGGAC AGGCGAAGGCTTCCT ; SEQ ID NO:693). HLA-A and RAS modified polyclonal pools were plated at a density of 0.5 to 2 cells per well in 96 well plates and cultured for 3 weeks. Single colonies were transferred into 2 or 3 replica 96-well plates. For HLA-A modified NCI-H441 cells, two replica plates were used. One plate was used to stain cells with the HLA-A3 specific antibody GAP.A3-PE and the other with anti-HLA-A2 specific antibody BB7.2-PE (BioLegend). Comparison of the staining allowed identification of clones with only the HLA-A3 allele knocked out, as NCI- H441 normally expresses both HLA-A2 and HLA-A3. HL-60 cells were stained with anti- HLA-A1 (clone 8.L.101) to select clones with HLA-A1 knocked out. For the RAS modified cells, genomic DNA was harvested from one of the plates using the Quick-DNA™ 96 Kit (Zymo Research), PCR amplified using Q5® Hot Start High-Fidelity 2X Master Mix (New England BioLabs), and Sanger sequenced to identify clones with the desired RAS mutation status. Targeted next generation sequencing was performed to confirm the mutation status of selected clones.

Co-cultures

COS-7 cells were plated in a T25 flask and transfected at 70% confluency. For each flask, 10 pg of plasmid DNA (1 : 1 ratio of HLA plasmidiiM^ plasmid, or control plasmid DNA, was mixed with 20 pL of P3000 Reagent in 250 pL of Opti-MEM (Thermo Fisher Scientific). In a separate tube, 20 pL of Lipofectamine 3000 Reagent was mixed with 250 pL of Opti-MEM. The contents of the two tubes were mixed and allowed to complex for 10 minutes at room temperature. Existing medium on pre-plated COS-7 cells was removed and fresh medium was added followed by the Lipofectamine-DNA mixture. Cells were harvested 24 hour post-transfection for co-culture after washing once with PBS, adding 1 mL 0.05% trypsin (Thermo Fisher Scientific) and incubation at 37°C for 5-10 minutes. Trypsin was quenched with serum-containing media and cells were counted. The co-culture was set up in 96-well flat-bottom tissue culture treated plates. To each well, the following components were combined: 50 pL of antibody diluted to the desired concentration in complete RPMI-1640 with a final IL-2 concentration of 100 IU/mL, 1 x 10⁴ COS-7s in 100 pL complete RPMI-1640, 5 x 10⁴ human T cells in 50 pL complete RPMI-1640. The co cultures were incubated for 24 hours at 37°C. The resultant supernatant was assayed for cytokines using a Human IFNy Quantikine and Human TNFa Quantikine ELISA Kits (both R&D Systems Bio-techne, Minneapolis, MN) according to the manufacturer’s instructions.

For co-cultures with pulsed cells, cells were peptide-pulsed in a 96-well plate by combining the specified number of target cells in 50 pL of serum-free RPMI-1640 media (for T2A3, SigM5, Hs 695T) or serum-containing RPMI-1640 media (for PBMCs, monocytes, iDCs, mDCs) with 50 pL of serially diluted peptide in serum-free RPMI-1640 media. Cells were incubated for 4 hours at 37°C, after which 5 x 10⁴ human T cells and antibodies were added in an additional 100 pL of serum-containing RPMI-1640 medium, with a final IL-2 concentration of 100 IU/mL. The co-cultures were incubated for 24 hours at 37°C, the cells pelleted by centrifugation at 500 g, and the cell-free supernatant was assayed for cytokines as described above. Cell viability was assayed by CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI) according to the manufacturer’s instructions. Cytotoxicity was calculated by taking the luciferase signal of a given well, subtracting the luciferase signal of the T cell only wells, and normalizing to the luciferase signal of the wells without scDb.

Co-cultures with other target cell lines were set up similarly, with target cells and human T cells combined as specified in the figure legends. For Luminex assays, IFNy, IL-2, TNFa, granzyme B, and perforin were measured using MILLIPLEX panels (MilliporeSigma, Burlington, MA).

Western Blots

COS-7 cells were transfected as above and harvested after 24 hours. Cells were pelleted, snap frozen, and re-suspended in radioimmunoprecipitation assay (RIP A) buffer (Thermo Fisher Scientific). 25 pg of cell lysate or approximately 0.5-1.5 pg of recombinant protein was run on mini-PROTEAN TGX stain-free gels (Bio-Rad, Hercules, California). Protein was transferred from the stain-free gel to a Trans-Blot Turbo Mini 0.2 pm PVDF membrane (Bio-Rad) using the Trans-Blot Turbo Transfer System (Bio-Rad). Membrane was blocked with blocking buffer (5% Bio-Rad Blotting-Grade Blocker Bio-Rad in lx TBST) on an orbital shaker (VWR) for 1 hour at RT. All antibody incubations were also performed using this blocking buffer. All incubations with primary antibodies were done at 4°C overnight (except for b-Actin as noted below). All incubations with secondary antibodies were done for 1 hour at RT. All washes were performed with lx TBST. KRAS was detected with mouse monoclonal antibody (mAb) F234 (Santa Cruz Biotechnology, Dallas, TX, at 1:500), followed by anti-mouse-HRP (Thermo Fisher Scientific, at 1:10k). HLA-A3 was detected with mouse mAb 7g7h8 (Abeam, at 1:500), followed by anti-mouse- HRP (Thermo Fisher Scientific, at 1 : 10k). Rab-7b was detected with rabbit anti-Rab7b mAb abl93360 (Abeam, at 1:500), followed by anti-rabbit-HRP (Thermo Fisher Scientific, at 1 : 10k). b-Actin was detected with rabbit mAb 13E5 conjugated to HRP (Cell Signaling Technology, Danvers, MA, at 1:3000) for 1 hour at RT. Recombinant proteins with 6xHIS tags were detected with anti-6x HIS tag rabbit polyclonal antibody ab9108 (Abeam, 1 : 1000), followed by anti-rabbit-HRP (Thermo Fisher Scientific, at 1 : 10k). Prior to imaging, the membranes were washed with ddH20, then imaged using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific) on a ChemiDoc XRS+ imager (Bio- Rad).

Viral transduction of cell lines

To generate luciferase-expressing NCI-H358 and HL-60 cell lines for in vivo experiments, NCI-H358 and HL-60 cells were transduced with RediFect Red-Fluc-GFP lentiviral particles (Perkin Elmer, Waltham, MA). Non-tissue culture-treated 48-well plates were coated with 250 pL of 20 pg/mL RetroNectin (Clontech) per well for 2 hours at RT, then blocked with 10% FBS for 1 hour at RT. Viral particles and 2 x 10⁵ target cells were added to each well in a total volume of 275 pL of cell culture media and subjected to centrifugation at 1200 x g for 1 hour at 20°C. Cell culture volumes were brought up to 500 pL with complete media. Cells were then incubated at 37°C for 3 days before exchanging media. Transduced cells were sorted based on the presence of GFP using FACSAria Fusion (BD Biosciences) 18 days after transduction.

Mouse xenograft model

Female NOD . ( 'g-Prkdc^scldll2rf^{m 1 ,l/V}/SzJ (NSG) mice at 6-10 weeks were acquired from the Johns Hopkins Sidney Kimmel Comprehensive Cancer Center Animal Resources Core and treated in compliance with a research protocol approved by the Johns Hopkins University Animal Care and Use Committee. Mice were maintained on an irradiated Uniprim rodent diet (Envigo, Indianapolis, IN). Littermate controls were used for all experiments. For the NCI-H358 intrasplenic model, 5 x 10⁵ luciferase-expressing NCI-H358 cells (CRISPR KRAS G12V-KI clone 1) or isogenic parental NCI-H358 clone ( KRAS G12C/WT) were inoculated into mouse spleens on day zero using sterile surgical techniques. Intraoperatively, two-week micro-osmotic pumps (model 1002, ALZET, Cupertino, CA) filled with V2-U scDb or isotype control (L2-U scDb) were placed intraperitoneally. Successful tumor inoculation was ensured by bioluminescence imaging one day later, followed by intravenous injection of 1 x 10⁷ human T cells via lateral tail vein. Bioluminescent imaging (IVIS imaging system) was performed using RediJect D-luciferin Ultra (Perkin Elmer) according to the manufacturer’s instructions (Perkin Elmer). Image analysis was performed using Living Image software. Individual luminescence measurements were normalized to the average fluorescence of the injection marker dye (745^ex/800^em) in the thoracic region.

For the HL-60 leukemia cell model, mice were inoculated intravenously with 1 x 10⁷ human T cells and 5 x 10⁵ luciferase-expressing parental HL-60 (NRAS WT/Q61 L) or isogenic control HL-60 ( NRAS Q61H/Q61H) via lateral tail vein injection. On day 1, mice were randomized based on luminescence to ensure similar pretreatment tumor burden. Two- week micro-osmotic pumps filled with L2-U scDb or isotype control scDb (V2-U scDb) that had been primed in 1 mL PBS overnight at 37°C were then placed intraperitoneally using sterile surgical techniques. Tumor growth was serially monitored by bioluminescent imaging.

Statistics

Statistical analyses were performed with Prism 8 (GraphPad Software, La Jolla, CA). Unless otherwise indicated, error bars represent the standard deviation of three technical replicates that were independently assembled. Error bars smaller than the symbols used to represent the mean of these replicates are not shown. Percent cytotoxicity of target cells for in vitro experiments was calculated as described above. For in vivo experiments, statistical significance was performed with an unpaired, two-tailed t-test with Bonferroni-Dunn correction for multiple comparisons. Example 4: Targeting a neoantigen derived from a common TP53 mutation

This Example describes an immunotherapeutic agent that targets a common TP53 mutation. Results from Example 2 were reanalyzed with additional samples and included in the following.

Results

The p53^R175H neoantigen is presented on the surface of cancer cells

The p53^R175H (aa 168-176, HMTEVVRHC; SEQ ID NO:l) and p53^WT (HMTEVVRRC; SEQ ID NO: 135) peptides were predicted on the NetMHCpan 4.0 server to bind HLA-A*02:01 at 5177.6 nM (rank 9.7%) and 7121.5 nM (11.6%), respectively. To provide experimental evidence of, and to quantify such presentation, peptides eluted from HLA molecules were analyzed in four different cell culture systems using a mass spectrometry (MS)-based method. First, the human HLA-A*02:01 and either full-length p53^R175H or p53^WT were co-expressed in monkey COS-7 cells. MS analysis of the peptides immunopurified with an anti-HLA antibody detected the p53^R175H peptide at approximately 700 copies per cell (Fig. 73 A, Table 17). Though relatively abundant amounts of the p53^R175H peptide were detected in pHLA complexes, the p53^WT peptide was not observed in pHLA complexes in transfected cells, despite equivalent amounts of p53^WT and p53^R175H total protein expression as assessed by Western blotting (Fig. 73B). Second, MS analysis was performed on three human cancer cell lines, KMS26, TYK-nu, and KLE, all of which harbor the p53^R175H mutation and carry an HLA-A*02:01 allele. The p53^R175H peptide was detected on all three cell lines, and at much lower levels than in the COS-7 cells in which the mutant TP53 and HLA genes were exogenously introduced (Fig. 73C, Table 17). Based on comparisons with heavy isotope labeled controls, it was estimated that there were 2.4, 1.3, and 1.5 copies of cell-surface p53^R175H/HLA-A*02:01 complexes on the cell surfaces of KMS26, TYK-nu, and KLE cell lines, respectively (Table 17).

Table 17. Quantitative assessment of the p53^R175H neoantigen peptide. The amount of p53^R175H neoantigen peptide (HMTEVVRHC) present in COS-7 cells transfected with HLA- A*02:01 and p53^R175Hor p53^WT, as well as cells lines that endogenously express HLA- A*02:01 and p53^R175H, were quantified using mass spectrometry. ^Corrected for peptide recovery (all cell lines) and transfection efficiency (COS-7).

Identification of scFv-expressing phage clones specific for the HLA-A*02: 01 -restricted p 53^R175H peptide and conversion to scDb format

To identify TCRm single-chain variable fragments (scFvs) selectively targeting mutant pHLA complexes, an scFv-displaying phage library was screened with an estimated complexity >1 x 10¹⁰. Positive selection against HLA-A*02:01 pHLA monomers containing the p53^R175H peptide were combined with negative selection against pHLA monomers containing the p53^WT and irrelevant peptides. Selected phage clones were amplified and assessed for binding to T2 cells presenting the mutant or wild-type (WT) peptide via flow cytometry (Fig. 74A).

Twenty-three phage clones with median fluorescence intensity (MFI) ratios of p53^R175H to p53^WT >4 were then converted to T cell-retargeting bispecific antibodies. This was achieved through linking each individual scFv to an anti-CD3r scFv (UCHT1) in a single-chain diabody (scDb) format (Fig. 74B). The scDb format was chosen after evaluating several previously described bispecific antibody formats, such as bispecific T-cell engager (BiTE), dual-affinity re-targeting antibody (DART), and diabody in pilot experiments assessing protein expression and in vitro T cell activation. The ability of scDbs to activate T cells was assessed by interferon-g (IFN-g) release after co-incubation with COS- 7 cells overexpressing HLA-A*02:01 and either full-length p53^R175H or p53^WT proteins. Two scDb clones, named H2-scDb and H20-scDb and derived from phage clones H2 and H20, respectively, showed the most potent and specific T-cell activation in the presence of p53^R175H/HLA-A*02:01 (Fig. 74C, Table 18). The specificity of these scDbs was further evaluated by titration enzyme-linked immunosorbent assay (ELISA). Both H2- and H20- scDb bound to p53^R175H/HLA-A*02:01 at low concentrations. At high concentrations, H20- scDb also bound to p53^WT/HLA-A*02:01, while H2-scDb did not bind to the WT pHLA complex even at very high concentrations of the scDb (Fig. 67A, Fig. 74D). H2-scDb was therefore chosen for further analysis. As assessed by surface plasmon resonance (SPR), the H2-scDb bound to p53^R175H/HLA-A*02:01 with a KD=86 nM, a k₀n of 1.76 x 10⁵M^_1 s ¹, and a koff of 1.48 x 10² s ¹ (Fig. 67B). The k_on of 1.76 x 10⁵M^_1 s ¹ suggested a lack of overall conformational change of the p53^R175H/HLA-A*02:01 upon binding. No detectable binding of the H2-scDb to p53^WT/HLA-A*02:01 was observed in the SPR experiments (Fig. 67B).

Attorney Docket No. 44807-0348WO1 / C15966_P15966-03

Table 18. Sequences of the top scFv clones from phage library selection.

Next, it was examined whether anti-CD3 arms of the scDb other than the original UCHT1, could influence the ability of H2 to induce T-cell activation. The H2-scFv was linked to a panel of commonly used anti-CD3r scFvs, including UCHT1, UCHTlv9, L2K-07, OKT3, and hXR32 (Fig. 75A, B). It was found that, among the anti-CD3r scFvs tested, UCHT1, which has the highest reported affinity (Table 19), activated T cells at the lowest p53^R175H peptide concentration when linked to the H2-scFv (Fig. 75C, Fig. 67C). H2- UCHTl-scDb (hereafter H2-scDb) was thus used for further experiments. Thermal stability of H2-scDb as measured by differential scanning fluorimetry showed a single melting temperature (T_m) at 69 °C, suggesting it being a stable molecule (Fig. 75D, E).

Table 19. Reported affinity of the anti-human CD38 scFvs used in the study. H2-scDb specifically recognizes cancer cells expressing the p53^R175H neoantigen

The ability of H2-scDb to recognize cancer cell lines expressing various levels of HLA-A*02:01 and having different p53 mutation status was next evaluated. H2-scDb elicited T-cell responses in a dose-dependent manner when T cells were co-cultured with three cell lines that expressed moderate to high levels of HLA-A*02:01 and harboring p53^R175H (KMS26, KLE, TYK-nu, as well as the cisplatin-resistant variant of TYK-nu) (Fig.

68A, B, Fig. 76A). This activation was noted even at very low (sub-nanomolar) concentrations of the bispecific antibody and the reactivity was strictly T cell and H2-scDb- dependent (Fig. 76B, C). The T-cell responses were polyfunctional, as indicated by the release of cytotoxic granule proteins granzyme B and perforin, cytotoxicity, and the production of cytokines IFN-g, tumor necrosis factor a (TNF-a), interleukin-2 (IL-2), and others (Fig. 68C, Fig. 76D-F). Clustering of T cells around tumor cells, leading to their lysis in the presence of H2-scDb, was also visualized by real-time live-cell imaging (Fig. 68D). The specificity of the bispecific antibody for both the p53^R175H peptide and HLA-A*02:01 was evident from the observation that much lower levels of IFN-g were induced by cells harboring a p53^R175H mutation but low levels of expression of HLA-A*02:01 (AU565 or SK- BR3) or by cells without p53^R175Hbut relatively high levels of HLA-A*02:01 expression (Fig. 68B, Fig. 77A).

The specificity of H2-scDb was validated using nine pairs of isogenic cell lines that differed with respect to HLA-A*02:01 expression or p53^R175H mutation (Fig. 69A). First, human HEK293FT (77^J53^,r//HLA-A*02:0 l ) or Saos-2 (77^J53"^!,///HLA-A*02:0 l ) cells were transfected with plasmids expressing either full-length p53^R175H or p53^WT. H2-scDb induced robust T-cell activation when co-cultured with both cell lines overexpressing p53^R175Hbut not with p53^WT-overexpressing or parental cells (Fig. 69B). Second, HLA-A*02:01 -encoding retrovirus was transduced into four cell lines (AU565, SK-BR-3, HuCCTl, CCRF-CEM) that harbored the p53^R175H mutation but had low levels of HLA-A*02:01 expression (Fig. 77B). Exogenous expression of HLA-A*02:01 in all four lines conferred T-cell activation by H2- scDb (Fig. 69C). Third, TP53 in KMS26, TYK-nu, and KLE cancer cell lines that carry endogenous p53^R175H were genetically disrupted using a CRISPR-based technology (Fig. 78A). T-cell activation, as assessed by IFN-g secretion, was reduced to control levels when TP53 was knocked out in all three cell lines (Fig. 69D). The cytotoxicity mediated by H2- scDb was similarly mitigated by TP53 knock-out (KO) in these cells (Fig. 69E, Fig. 78B).

Overall structure of the H2-Fab p53R175H/HLA-A *02:01 ternary complex

To understand the structural basis for the high specificity of the H2 clone for p53^R175H/HLA-A*02:01, H2 was converted into full-length IgG (H2-IgG) and confirmed that binding specificity was preserved in this format (Fig. 79A). The H2-IgG was then digested into an antigen-binding fragment (H2-Fab) with papain (Fig. 79B). The H2-Fab- p53^R175H/HLA-A*02:01 complex was purified (Fig. 79C, D) and its crystal structure was determined by molecular replacement and refined to 3.5 A resolution (PDB ID 6W51, Table 20). There were four H2-Fab and four p53^R175H/HLA-A*02:01 per asymmetric unit (Fig. 70A, B). All four H2-Fab were firmly positioned on the p53^R175H/HLA-A*02:01 with a pairwise root-mean-square deviation (rmsd) range of 0.27 to 0.45 A for 382 to 419 C^alpha carbon’s, as calculated by PyMOL (Table 21). The total buried surface area of the H2-Fab- p53^R175H/HLA-A*02:01 interface was 1173 A², with roughly equal contributions from heavy and light chains (644 A² and 529 A², respectively, Table 22). Although the entire structure was refined to a resolution of 3.5 A, particularly clear electron densities were observed for the p53^R175H neoantigen, the complementarity-determining regions (CDRs) of the H2-Fab, and the FiLA-A*02:01 (Fig. 70C, D).

Table 20. X-ray Crystallography data collection and refinement statistics. p53^R175H/HL A- A * 02 : 01-Fab H2

Data Collection

Diffraction source NSLS-II X17-ID-2 Wavelength (A) 0.979321 Temperature (K) 100 Detector Dectris EIGER X 16M Space group P 12₁1 a, b, c (A) 113.3, 123.7, 136.9 a, b, g (°) 90, 100.4, 90

Resolution range (A) 30.37-3.53 (3.66-3.53) Total no. of reflections 104,474 (9,774) No. of unique reflections 43,734 (4,273) Completeness (%) 95.3 (89.2) Redundancy 2.4 (2.3)

Refinement

Resolution range (A) 30.38-3.53 (3.62-3.53)

No. of reflections, working set 41,530

Rwork/Rfree 0.20/0.28 (0.30/0.35)

No. of non-H atoms

MHC (HLA-A* 02:01 + b2ih) 3,078 p53^R175H peptide 75

Fab H2 Heavy Chain 1,667

Fab H2 Light Chain 1,652

Total of non-H atoms 6,472

Rm.s. deviations Bonds (A) 0.009 Angles (°) 1.66

Wilson B-factor (A²) 62 Average B factors (A²)

MHC (HLA-A* 02 : 01 + (32m) 74 p53^R175H peptide 56

Fab H2 Heavy Chain 59

Fab H2 Light Chain 63

Total average B factor 63

Ramachandran (%)

Favorable 95.2

Allowed 3.8

Outlier 1.0

*Values in parentheses are for highest-resolution shell. All atoms refer to non-H atoms.

Table 21. Pairwise rmsd of the four H2-Fab in the asymmetric unit. The root-mean square deviation (rmsd) and number of C^alpha carbons were calculated by PyMOL (v2.2.3, Schrodinger, LLC, New York, NY). For chain reference see PDB ID 6W51.

Table 22. Structural comparison of H2-Fab-p53^R175H/ HLA-A*02:01 with various TCR and Fab antibody -pHL A. Total bonds were calculated using a 4 A cutoff which includes both hydrogen bonds and van der Waals interactions as calculated using CONTACTS in the CCP4 suite. PDB, Protein Data Bank; BSA, buried surface area; a, TCRa chain; b, TCRP chain; H, VH domain; L, VL domain; pep, HLA presented peptide.

Binding of the p53^R175H peptide to HLA-A *02:01

The p53^R175H peptide (HMTEVVRHC; SEQ ID NO: 1) occupied the binding cleft al- a2 of HLA-A* 02:01, burying a solvent accessible surface area of 870 A², slightly larger than other peptide/HLA-A*02:01 complexes (Fig. 71 A, B, Fig. 80A) and with the C-terminal arginine at position 7 (Argl74) and mutant histidine at position 8 (Hisl75) pointing up, out of the groove. In contrast, the N-terminus of the peptide is situated deep within the peptide binding cleft, anchored by multiple residues in the HLA-A*02:01 (Fig. 71 A, B, Fig. 80A). The anchor residues of the peptide, a methionine at position 2 (P2. Metl69) and a cysteine residue at position 9 (P9, Cysl76) (Fig. 80B), departed from the canonical anchor residues — leucine at P2 and valine or leucine at P9. Peptides that bind to HLA-A*02:01 through either a methionine at P2 or a cysteine at P9 have been reported, but not both (Webb et al., ./. Biol. Chem. 279:23438-23446 (2004); and Ataie et al., J. Mol. Biol. 428:194-205 (2016)). Based on alignments with structures of other HLA-A*02:01 peptides in complex with TCR or TCRm, the unconventional anchoring of p53^R175H did not result in drastic peptide conformational change or positioning (Fig. 80C, D).

Structural basis for the recognition of p53^R17m/HLA-A*02:01 by the H2-Fab

The recognition of the HLA-A*02:01 by the H2-Fab was mediated by all six CDRs. There were a total of 79 contacts, with a cutoff of 4 A, between the H2-Fab CDRs and the al and a2 of HLA-A* 02:01, with the light chain contributing to 61% of those contacts (Table 22 ). The H2-Fab buried a solvent accessible surface area of 818 A² within the HLA, of which 427 A² were contributed by the light chain and 391 A² by the heavy chain (Table 22). In contrast, only four of the six H2-Fab CDRs (HI, H2, H3 and L3) interacted with the p53^R175H peptide. Overall, the H2-Fab made 36 contacts with the p53^R175H neoantigen, including five hydrogen bonds and numerous van der Waals interactions. Hisl75 made 47% of all direct contacts with the H2-Fab. The CDR-H1, H2, and H3 of the heavy chain and CDR-L3 of the light chain formed a cage-like configuration around the C-terminus of the p53^R175H peptide, trapping Argl74 and Hisl75 into position by providing a stable interaction (Fig. 71C). The imidazole side chain of Hisl75 was anchored by a hydrogen bonding network with Asp54 (CDR-H2) and Tyr94 (CDR-L3) (Fig. 71C, Fig. 81). Tyr52 (CDR-H2) acted as a ceiling and capped the cage-like structure around Hi si 75 by forming p-p interactions (Fig. 71C, Fig. 81).

Viewed from the axis of the C-terminus to the N-terminus of the p53^R175H peptide, the CDRs were arranged in the order H2, HI, L3, H3, LI, L2 (Fig. 70E, F, G). The docking angle of the H2-Fab to the peptide within the HLA groove is 36° (Fig. 70G, H). This orientation angle was quite different from that of most previously described TCRs or TCRm antibodies to pHLA complexes, in which the axis of the peptide is almost perpendicular to the axis defined between the disulfide bonds of the Vi7a to Vi-i/b chains (Fig. 82).

Assessing candidate cross-reactive peptides

One of the major challenges confronting new immunotherapeutic antibodies is off- target binding, which can result in toxicity to normal cells. Scanning mutagenesis was employed to identify peptides in the human proteome to which H2-scDb might cross-react.

A peptide library was generated by systemically substituting amino acids at each position of the target p53^R175H peptide (HMTEVVRHC; SEQ ID NO: 1) with each of the remaining 19 common amino acids. T2 cells loaded with each of the 171 variant peptides were then used to assess T-cell activation by measuring IFN-g release following incubation with T cells and H2-scDb (Fig. 7 ID). In congruence with the X-ray structural analysis, any changes in P8, where the mutant histidine residue lies, and any change in P7, which is encased with P8 by the CDR loops, abolished recognition of the peptide. Peptides with substitutions at these positions retained their ability to bind to HLA-A*02:01 (Fig. 83 A), but not to the H2-scDb. Other non-anchor residues at positions 3-6 also highly favored the parental amino acids present in the target peptide. This recognition pattern is illustrated as a Seq2Logo graph (Fig. 71E).

Next, a nonamer binding motif, x-[AILMVNQTC]-[ST]-[DE]-[IV]-[IMVST]-R-H- [AILVGHSTYC] (SEQ ID NO: 197), was generated using 20% target peptide reactivity as a cutoff for permissive amino acids at each position (Fig. 83B). A search of this motif in the UniProtKB human protein database using ScanProsite yielded 3 homologous peptides from STAT2 (PLTEIIRHY; SEQ ID NO: 185), VPS13A (LQSEVIRHY; SEQ ID NO: 186), and ZFP3 (QNSEIIRHI; SEQ ID NO: 187) (Table 9). None of these 3 peptides were predicted to be potent binders of HLA-A*02:01 by NetMHCpan 4.0 (% rank all >2.0) and had lower predicted binding affinity than the parental p53^R175H peptide (Table 9). However, to experimentally exclude the possibility of cross-reactivity, T2 cells were pulsed with each of these peptides. H2-scDb activated T cells only in the presence of T2 pulsed with the p53^R175H peptide (Fig. 7 IF). Additionally, COS-7 cells were co-transfected with expression plasmids for HLA-A*02:01 and full-length STAT2 or ZFP3; VPS13A was not tested due to its large size (>3000 aa). Again, no T-cell activation was detected in the co-culture assay with COS-7 cells expressing the two proteins containing the candidate cross-reactive peptides (Fig. 83C, D).

Antitumor activity of the H2-scDb in vivo

To determine whether H2-scDb could control tumor growth in vivo , KMS26 multiple myeloma cells were engrafted into NOD-SCTD-J/irg (NSG) mice through intravenous injection, establishing widespread, actively growing cancers throughout the body. Two models were used to assess the effects of the H2-scDb in combination with human T cells engrafted in these mice (Fig. 84A). In an early treatment model, mice were randomized based on luminescence quantification of tumor burden and H2-scDb was subsequently administered through continuous intraperitoneal infusion pumps at 0.3 mg/kg/day, starting one day after tumor inoculation. The pumps were able to maintain significant plasma concentrations of scDb for two weeks (Fig. 84B). An irrelevant isotype scDb was administered to mice in parallel as control. H2-scDb markedly suppressed the growth of parental KMS26 tumors (Fig. 72A). In contrast, the H2-scDb had no effect on KMS26 tumors in which the TP53 gene had been disrupted using CRISPR (Fig. 72A). In the second model, mice were randomized 6 days after tumor inoculation. The H2-scDb was administered at two doses (0.15 and 0.3 mg/kg/day). Both doses resulted in major tumor regressions and were well-tolerated as assessed by the absence of significant changes in body weight (Fig. 72B, Fig. 84C). No treatment effect of H2-scDb was observed in the absence of human T cells, supporting the T cell-dependent nature of H2-scDb (Fig. 84D).

Together these results demonstrate that highly specific bispecific antibodies can be generated against pHLA complexes resulting from common mutations occurring in cancer cells. The format and configuration of the bispecific antibodies developed here are highly specific and sensitive scDbs against protein products containing mutations occurring in cancer cells.

Materials and Methods

Cell lines and primary T cells

COS-7, RPMI 6666, T2 (174 x CEM.T2), Raji, HH, AU565, SK-BR-3, KLE,

HCT116, SW480, NCI-H441, Saos-2, and CCRF-CEM cells were purchased from American Type Culture Collection (ATCC, Manassas, VA). KMS26, TYK-nu, and HuCCTl were purchased from Japanese Collection of Research Bioresources Cell Bank (JCRB, Osaka, Japan). SigM5 was obtained from DSMZ (Braunschweig, Germany). HEK293FT was obtained from Invitrogen (Thermo Fisher Scientific, Waltham, MA). T2, Raji, Jurkat, HH, AU565, NCI-H441, TOV-112D, CCRF-CEM, KMS26, TYK-nu, TYK-nu.CP-r and HuCCTl were cultured in RPMI- 1640 (ATCC, 30-2001) with 10% FBS (GE Healthcare, SH30070.03) and 1% Penicillin-Streptomycin (Thermo Fisher Scientific, 15140163). RPMI 6666 was cultured in RPMI-1640 with 20% FBS and 1% Penicillin-Streptomycin. COS-7, SK-BR-3, HCT116, SW480, and Saos-2 were cultured in McCoy’s 5A modified media (Thermo Fisher Scientific, 16600108) with 10% FBS and 1% Penicillin-Streptomycin.

SigM5 was cultured in IMDM (Thermo Fisher Scientific, 12440061) with 20% FBS and 1% Penicillin-Streptomycin. HEK293FT was cultured in DMEM (high glucose, pyruvate, Thermo Fisher Scientific, 11995065) with 10% FBS, additional 2 mM GlutaMAX (Thermo Fisher Scientific, 35050061), 0.1 mM MEM non-essential amino acids (Thermo Fisher Scientific, 11140050), 1% Penicillin-Streptomycin, and 500 pg/mL Geneticin (Thermo Fisher Scientific, 10131027). PBMCs were isolated from leukapheresis samples (Stem Cell Technologies, Vancouver, BC) by standard density gradient centrifugation with Ficoll Paque Plus (GE Healthcare, 17-1440-03). T cells were expanded from PBMCs with addition of the anti-human CD3 antibody (OKT3, BioLegend, San Diego, 317347) at 15 ng/mL for three days. T cells were cultured in RPMI-1640 with 10% FBS, 1% Penicillin-Streptomycin, 100 IU/mL recombinant human IL-2 (aldesleukin, Prometheus Therapeutics and Diagnostics, San Diego, CA), and 5 ng/mL recombinant human IL-7 (BioLegend, 581908). In general, T cells from at least two different donors were tested in in vitro assays. All cells were grown at 37°C in 5% CO2 with humidification.

Detection of neoantigen peptide

HLA-A*02:01 restricted p53^R175H peptide was directly detected and quantified in human cancer cells carrying p53^R175H mutations through MANA-SRM in COS-7 cells transfected with HLA-A*02:01 and p53^R175H and in human cancer cells carrying p53^R175H mutations and expressing HLA-A*02:01. In particular, the dual -reduction approach described in MANA-SRM was critical for this detection because a cysteine and a methionine coexist in the p53^R175H peptide. One hundred femtomole heavy-isotope labeled p53^R175H peptide HMTEVVRHC (SEQ ID NO: 1) and p53^WT peptide HMTEVVRRC (SEQ ID NO: 135; New England Peptide Inc, Gardner, MA) were spiked into each sample before the assay. The MANA-SRM assays were performed at Complete Omics Inc. (Baltimore, Maryland).

Peptides and monomers

All peptides were synthesized at a purity of >90% by Peptide 2.0 (Chantilly, VA) or ELIM Biopharm (Hayward, CA), except for the positional scanning library, where crude peptides were used. Peptides were resuspended in dimethylformamide at 10 mg/mL and stored at -20°C. Biotinylated pHLA monomers were synthesized by Fred Hutchinson Cancer Research Center Immune Monitoring Lab (Seattle, WA). Monomers were confirmed to be folded prior to selection by performing an ELISA using W6/32 antibody (BioLegend,

311402), which recognizes only folded HLA. Phage display library construction

The scFv-bearing phage library used in this study has been described elsewhere (see, e.g., Miller et al., J Biol. Chem. 294:19322-19334 (2019)). Briefly, oligonucleotides were synthesized by GeneArt (Thermo Fisher Scientific) using trinucleotide mutagenesis (TRIM) technology to diversify complementarity-determining region (CDR)-L2, CDR-L3, CDR-H1, CDR-H2, and CDR-H3. A FLAG (DYKDDDDK; SEQ ID NO: 190) epitope tag was placed immediately downstream of the scFv, which was followed in-frame by the full-length Ml 3 pill coat protein sequence. The total number of transformants obtained was determined to be 3.6 x 10¹⁰.

Selection of mutant pHLA specific phage clone

Phage clones bearing scFvs specific to p53^R175H/HLA-A*02:01 pHLA were identified using an approach described elsewhere (see, e.g. , Skora etal , 2015 PNASA 12:9967-72). One pg of biotinylated HLA-A*02:01 pHLA monomer complexes were conjugated to 25 pL of M-280 streptavidin magnetic Dynabeads (Thermo Fisher Scientific, 11206D). During the enrichment phase (Round 1), phages were negatively selected with a mixture of unconjugated Dynabeads and free streptavidin protein (RayBiotech, Norcross, GA, 228- 11469). After negative selection, supernatant containing unbound phages were transferred for positive selection using 1 pg of p53^R175H/HLA-A*02:01 pHLA. Beads were then washed and phages were eluted to infect mid-log-phase SS320 bacteria, with the addition of M13K07 helper phages (multiplicity of infection of 4). Bacteria were then grown overnight at 3 CPC for phage production and the phages were precipitated the next morning with PEG/NaCl.

During the selection phase (Rounds 2-5), phages from the previous round were subjected to two stages of negative selection: 1) against cell lines without p53^R175H/HLA- A*02:01 (RPMI 6666, Jurkat, Raji, SigM5, HH, T2, and NCI-H441) and 2) against p53^WT/HLA-A*02:01 pHLA, unrelated HLA-A*02:01 pHLA, and free streptavidin. For negative selection using cell lines, phages were incubated with a total number of 0.5-1 x 10⁷ of cells at 4°C overnight. After negative selection, beads were isolated and unbound phages were transferred for positive selection by incubating with 1 pg (Round 2), 0.5 pg (Round 3), or 0.25 pg (Round 4, 5) of p53^R175H/HLA-A*02:01 pHLA. Phages were then eluted and amplified by infecting SS320 as described above. After five rounds of selection, SS320 cells were infected with a limiting dilution of the enriched phages. A total of 190 individual colonies of SS320 were picked and phage DNA was PCR amplified by primers flanking the CDRs (Forward:

GGC CAT GGC AGAT ATTC AG A (SEQ ID NO: 198), Reverse:

CCGGGCCTTTATCATCATC (SEQ ID NO: 199)) using Q5 Hot Start High-Fidelity 2X Master Mix (New England BioLabs, M0494L) and Sanger sequenced by GENEWIZ (South Plainfield, NJ). Sequences flanking the CDRs were trimmed using DNA Baser Sequence Assembler v4 (Arges, Romania) and the sequences spanning the CDRs were clustered using the CD-HIT Suite. Colonies containing unique phage clones were selected and grown overnight in 400 pL of media in deep 96-well plates (Thermo Fisher Scientific, 278743) with the addition of M13K07 helper phages. Bacteria were pelleted the next day and the phage laden supernatants were used for downstream analysis.

Peptide pulsing

For peptide pulsing, T2 cells were washed with serum-free RPMI-1640 media before incubation at 0.5-1 x 10⁶ cells per mL in serum-free RPMI-1640 containing peptides at the specified concentration of 2 hours at 37°C. For experiments assessed using flow cytometry, human b2M (ProSpec, East Brunswick, NJ, PRO-337) at 10 pg/mL was added with the peptides and specified in the figure legends of such experiments.

Flow cytometry

Phage staining of peptide-pulsed T2 cells was performed with 50 pL phage supernatant on ice for 1 hour, followed by staining with 1 pg of rabbit anti -Ml 3 antibody (Novus Biologicals, NB100-1633), and anti-rabbit-PE (BioLegend, 406421). HLA-A*02 staining was performed by staining cells with fluorescently labeled anti-human HLA-A*02 (BB7.2, BioLegend, 343308) or mouse isotype IgG2b, k (BioLegend, 402206). Stained cells were analyzed using an LSRII flow cytometer (Becton Dickinson, Mansfield, MA) or an iQue Screener (IntelliCyt, Albuquerque, NM).

ELISAs

Streptavidin-coated, 96-well plates (R&D Systems, Minneapolis, MN, CP004) were coated with 50 ng of biotinylated HLA-A*02:01 pHLA monomers in 50 pL of blocking buffer (PBS with 0.5% BSA, 2 mM EDTA, and 0.1% sodium azide) or 25 ng of recombinant human CD3e/5 (Aero Biosystems, DE, CDD-H52W4) at 4°C overnight. Plates were washed with IX TBST (TBS + 0.05% Tween-20) using a BioTek 405 TS plate washer (BioTek, Winooski, VT). Serial dilutions of scDb or IgG was incubated on the plate for 1 hour at RT and washed. For scDbs, the plate was then incubated with 1 pg/mL recombinant protein L (Thermo Fisher Scientific, 77679) for 1 hour at RT, washed, followed by incubation with anti-protein L HRP (1 : 10000, Abeam, ab63506) for 1 hour at RT. For IgG, the plate was incubated with anti-human IgG HRP (1 : 1000, Thermo Fisher Scientific 62-8420) for 1 hour at RT. Plates were washed, 50 pL of 3,3',5,5'-Tetramethylbenzidine (TMB) substrate (BioLegend, 4211101) was added to each well, and the reaction was quenched with 50pl 2N sulfuric acid (Thermo Fisher Scientific). Absorbance at 450 nm was measured with a Synergy HI Multi-Mode Reader (BioTek). scDb production scDbs were produced by cloning gBlocks (IDT, Coralville, Iowa) encoding each of the variants in the format (from N- to C-terminus): IL-2 signal sequence, anti-pHLA variable light chain (VL), GGGGS (SEQ ID NO:200) short linker, anti-CD3 variable heavy chain (VH), (GGGGS)3 (SEQ ID NO:201) long linker, anti-CD3 VL, GGGGS (SEQ ID NO:200) short linker, anti-pHLA VH, and 6 x HIS tag into linearized pcDNA3.4 vector (Thermo Fisher Scientific, A14697). The proteins were expressed by the Eukaryotic Tissue Culture Core Facility of Johns Hopkins University. Briefly, 1 mg of plasmid DNA was transfected with polyethylenimine (PEI) at a ratio of 1:3 into 1 L of FreeStyle 293-F cells at a concentration of 2-2.5 x 10⁶ cells per mL and the transfected cells incubated at 37°C. Five days after transfection, culture media was collected and filtered through a 0.22-pm unit. The scDbs were purified using HisPur Ni-NTA Resin (Thermo Fisher Scientific, 88222) and desalted into PBS pH 7.4 or 20 mM Tris pH 9.0, 150 mM NaCl using 7k MWCO Zeba Spin desalting columns (Thermo Fisher Scientific, 89890). Proteins were quantified using a 4 - 15% Mini-PROTEAN TGX gel (Bio-Rad, Hercules, CA, 4568085) and/or NanoDrop (Thermo Fisher Scientific). Alternatively, the scDb proteins were produced by GeneArt (Thermo Fisher Scientific) in Expi293s, purified with a HisTrap column (GE Healthcare, 17- 5255-01) followed by size exclusion chromatography with a HiLoad Superdex 20026/600 column (GE Healthcare, 28989335). Analytic chromatography was performed using TSKgel G3000SWxl column (TOSOH Bioscience, Tokyo, Japan) using a running buffer of 50 mM sodium phosphate and 300 mM sodium chloride at pH 7, at a flow rate of 1.0 mL/minute.

Surface plasmon resonance affinity measurements of p53^R175H/HLA-A *02:01 and H2-scDb interaction

Biotinylated p53^R175H/HLA-A*02:01, p53^WT/HLA-A*02:01, and H2-scDb binding experiments were performed at 25°C using a Biacore T200 SPR instrument (GE Healthcare). Approximately 100-110 response units (RU) of biotinylated p53^R175H/HLA-A*02:01 and p53^WT/HLA-A*02:01 were captured in flow cells (Fc) 2 and 4, respectively, using a streptavidin chip. Single-cycle kinetics were performed by injecting increasing concentrations (3, 12, 50, 200, and 800 nM) of purified H2-scDb which was flowed over Fc 1-4. Binding responses for kinetic analysis were both blank- and reference- subtracted. Both binding curves were fit with a 1 : 1 binding model using Biacore Insight evaluation software.

Differential Scanning Fluor imetry Thermal stability of the H2-scDb was evaluated by a differential scanning fluorimetry

(DSF) assay which monitor the fluorescence of a dye that binds to the hydrophobic region of a protein as it becomes exposed upon temperature induced denaturation. Reaction mixture (20 pL) was set up in a white low-profile 96-well, unskirted polymerase chain reaction plate (BioRad, MLL9651) by mixing 2 pL of purified H2-scDb at a concentration of 1 mg/mL (final concentration ~2 pg) with 2 pL of 50X SYPRO orange dye (Invitrogen, S6650, 5X final concentration) in PBS, pH 7.4. The plate was sealed with an optical transparent film and centrifuged for 1,000 x g for 30 seconds. Thermal scanning was performed from 25 to 100 °C (1 °C/minute temperature gradient) using a CFX9 Connect real-time polymerase chain reaction instrument (BioRad). Protein unfolding/melting temperature T_m was calculated from the maximum value of the negative first derivative of the melt curve using CFX Manager software (BioRad).

CRISPR-mediated knockout of TP 53

The Alt-R CRISPR system (IDT) was used to knock out the TP53 gene from KMS26, TYK-nu, and KLE cell lines. CRISPR-Cas9 crRNAs targeting TP53 exon 3 (p53-5: CCCCGGACGATATTGAACAA (SEQ ID NO: 191) or p53-6:

CCCCTTGCCGTCCCAAGCAA (SEQ ID NO: 202)) as well as CRISPR-Cas9 tracrRNA were resuspended at 100 mM with Nuclease-Free Duplex Buffer. The crRNAs and tracrRNA were duplexed at a 1:1 molar ratio for 5 minutes at 95°C followed by cooling down slowly to RT according to the manufacturer’s instructions. The duplexed RNA was then mixed with Cas9 Nuclease at a 1.2:1 molar ratio for 15 minutes. A total of 40 pmols of the Cas9 RNP complexed with TP 53 gRNA were mixed with 2 x 10⁵ cells in 20 pL of OptiMEM. This mixture was loaded into a 0.1 cm cuvette (Bio-Rad, 1652089) and electroporated at 120V and 16 ms using an ECM 2001 (BTX, Holliston, MA). Cells were transferred to complete growth medium and cultured for 7 days. Single cell clones were established by limiting dilution and genomic DNA was harvested using a Quick-DNA 96 Kit (Zymo Research, Irvine, CA, D3012). A region flanking the CRISPR cut site was PCR amplified (forward primer: GCTGCCCTGGTAGGTTTTCT (SEQ ID NO:203), reverse primer: GAGACCTGTGGGAAGCGAAA (SEQ ID NO:204)) and Sanger sequenced to select for clones with the desired TP53 status.

Immunoblotting analysis

Cells were lysed in cold RIPA buffer (Thermo Fisher Scientific, 89901) supplemented with protease inhibitor cocktail (Thermo Fisher Scientific, 87785). Protein concentration was determined using a BCA assay (Thermo Fisher Scientific, 23227). Equal amounts of total protein (20-50 pg) were loaded in each lane of a 4-15% Mini-PROTEAN TGX gel (Bio-Rad, 4568085) and transferred to polyvinylidene difluoride membranes after electrophoresis. The membranes were incubated with appropriate primary antibodies (anti- ex His tag, 1:2000, Abeam, ab9108; p53 [DO-1], 1:1000, Santa Cruz, sc-126; STAT2, 1:1000, Thermo Fisher Scientific, 44-362G; ZFP3, 1:1000, Thermo Fisher Scientific, PAS- 62726; b-actin [13E5], 1:1000, Cell Signaling Technology, 5125S; b-actin [8H10D10], 1:1000, Cell Signaling Technology, 3700S) and species-specific HRP-conjugated secondary antibodies (1 :5000-10000). Signal was detected by a ChemiDoc MP chemiluminescence system (Bio-Rad). Transfection of cell lines gBlocks (IDT) encoding HLA and target proteins were cloned into pcDNA3.1 or pcDNA3.4 vectors (Thermo Fisher Scientific, V79020, A14697). COS-7, HEK293FT, and Saos-2 cells were transfected at 70-80% confluency using Lipofectamine 3000 (Thermo Fisher Scientific, L3000015) and incubated at 37°C overnight. A total of 15 pg and 30 pg plasmid (1 : 1 ratio of HLA plasmid/target protein plasmid in co-transfections) was used for T25 and T75 flasks, respectively.

Viral transduction of cell lines

HLA-A*02:01-encoding retrovirus was produced using the MSCV retroviral expression system (Clontech, Mountain View, CA, 634401). In brief, a gBlock encoding HLA-A*02:01-T2A-GFP (IDT) was cloned into the pMSCVpuro retroviral vector by HiFi DNA assembly (New England Biolabs, Ipswich, MA, E2621L). The pMSCVpuro-HLA- A*02:01-T2A-GFP plasmid was then co-transfected with a pVSV-G envelope vector into the GP2-293 packaging cell line. Viral supernatant was harvested 48 hours after transfection and concentrated 20-fold using Retro-X Concentrator (Clontech, 631456). RediFect Red-Fluc- GFP lentivirus particles (Perkin Elmer, Waltham, MA, CLS960003) was used for generating luciferase-expressing cell lines. NucLight green lentivirus (Essen Bioscience, Ann Arbor,

MI, 4624) was used to generate TYK-nu cell lines with nuclear GFP expression.

For transduction, non-tissue culture-treated 48-well plates were coated with 200 pL of 10 pg/mL RetroNectin (Clontech, T100B) per well overnight at 4°C and blocked with 10% FBS for 1 hour at RT. Viral particles and 2 x 10⁵ target cells were added to each well in a total volume of 500 pL cell culture media and spun at 2000 x g for 1 hour then incubated at 37°C. Selection with 1 pg/mL puromycin (Thermo Fisher Scientific, A1113803) began three days later. Transduced cells were sorted based on presence of GFP using FACSAria Fusion (BD Biosciences, San Jose, CA) 10-14 days after transduction.

In vitro scDb co-incubation assays

To each well of a 96-well flat-bottom plate, the following components were combined in a final volume of 100 pL RPMI-1640 with 10% FBS, 1% Penicillin- Streptomycin, and 100 IU/mL IL-2: scDb diluted to the specified concentration, 5 x 10⁴ human T cells, and 1-5 x 10⁴ target cells (COS-7, T2, or other tumor cell lines). The effector to target cell ratio is specified in the figure legend for each experiment. The co-culture plate was incubated for 20 hour at 37°C and conditioned media was assayed for cytokine and cytotoxic granule protein secretion using the Human IFN-g Quantikine Kit (R&D Systems, Minneapolis, MN, SIF50), Human IFN-g Flex Set Cytometric Bead Array (BD, 558269), or the MILLIPLEX Luminex assays (Millipore Sigma, HS T CM AG28 SPMX 13 , HCD8MAG- 15K) read on the Bioplex 200 platform (Bio-Rad). Cytotoxicity was assayed by CellTiter- Glo Luminescent Cell Viability Assay (Promega, Madison, WI, G7571), Bio-Glo Luciferase Assay (Promega, G7941), or Steady-Glo Luciferase Assay (Promega, E2510) per manufacturer’s instructions. For CellTiter-Glo assays, percent cytotoxicity was calculated by subtracting the luminescence signal from the average of the T cell only wells and normalizing to the no scDb condition: 1 - (scDb well - T cell only)/(no scDb well - T cell only) x 100. For Bio-Glo assays, percent cytotoxicity was calculated by normalizing luminescence signal to the no scDb condition: 1 - (scDb well)/(no scDb well) x 100.

Real-time live-cell imaging

A total of 1 x 10⁴ NucLight Green-labeled target cells were plated in each well of a 96-well flat bottom plate and allowed to attach for 4 hours before adding 2 x 10⁴ T cells and scDb at the indicated concentrations. Each condition was plated in triplicate. Plates were imaged every 6 hours using the IncuCyte ZOOM Live-Cell analysis system (Essen Bioscience) for a total of 120 hours. Four images per well at 10X zoom were collected at each time point. The number of GFP positive objects per mm² in each well was quantified using the green fluorescence channel.

Expression, purification and refolding of p53^R175H/HLA-A *02:01

Plasmids for HLA-A*02:01 and b2M were received from the NIH Tetramer Facility (Atlanta, GA) and separately transformed into BL21(DE3) cells. Each was expressed in inclusion bodies using auto-induction media. Purification of the HLA-A*02:01 and b2M inclusion bodies was achieved with a series of detergent washes followed by solubilization with 8 M urea. Refolding of the HLA-A*02:01, b2M, and mutant p53^R175H peptide was performed. Briefly, solubilized HLA-A*02:01 and b2M were combined in a refolding buffer containing 100 mM Tris pH 8.3, 400 mM L-arginine, 2 mM EDTA, 5 mM reduced glutathione, 0.5 mM oxidized glutathione, 2 mM PMSF, and 30 mg of the mutant p53^R175H peptide (aa 168-176, HMTEVVRHC; SEQ ID NO:l) dissolved in 1 mL of DMSO. The resultant solution was stirred at 4°C for 2 days, with two further additions of HLA-A*02:01 on day 2, concentrated to 10 mL and purified by size exclusion chromatography on a HiLoad 26/60 Superdex 75 Prep grade column (GE Healthcare, 28989334). For incubation with the H2-Fab, purified pHLA-A*02:01 was concentrated to ~l-3 mg/mL and stored at -80°C until use.

Production of the H2-Fab antibody fragment

The light chain (LC) and heavy chain (HC) variable region sequences of H2 scFv were grafted onto the respective constant chains of trastuzumab and separately cloned into a pcDNA3.4 vector (Thermo Fisher Scientific, A14697). Both chains were preceded by a mouse IgKVIII signal peptide. Before large-scale expression of full-length antibody, optimization of the LC:HC DNA ratio for transfection was performed to determine optimal recombinant protein yields. For a 1 L expression, a total of 50 pg of purified plasmids (1 : 1 LC:HC ratio) were transfected with PEI at a ratio of 1 :3 into Freestyle 293-F cells at a concentration of 2-2.5 x 10⁶ cells per mL and incubated at 37°C for 7 days. The media was harvested via centrifugation, filtered through a 0.22-pm unit and the full-length antibody was purified via protein A affinity chromatography on a HiTrap Mab Select™ SuRe™ column (GE Healthcare, 29-0491-04). Full-length antibody was eluted using a linear gradient of 0- 100 mM sodium citrate, pH 3.5. The protein A fractions containing pure H2 antibody were pooled, quantified by SDS-PAGE gel electrophoresis and dialyzed into 20 mM sodium phosphate buffer, pH 7.0, 10 mM EDTA.

For generation of H2-Fab fragments, ~l-3 mg of full-length antibody was mixed with 0.5 mL of a 50 % Immobilized Papain slurry (Thermo Fisher Scientific, 20341) pre-activated with digestion buffer (20 mM sodium phosphate buffer, pH 7.0, 10 mM EDTA) containing 20 mM cysteine-HCl. The mixture was incubated at 37°C overnight with constant shaking at 200 rpm. The H2 antibody digest was separated from the immobilized resin by a gravity resin separator and washed with 10 mM Tris-HCl, pH 7.5. Newly generated H2-Fab fragments were further purified by cation-exchange chromatography using a Mono-S column (GE Healthcare, 17516801) and eluted using a linear gradient of 0-500 mM NaCl. The H2-Fab fragments were concentrated, mixed with equimolar p53^R175H/HLA- A*02:01 and incubated at 4°C overnight. The H2-Fab-p53^R175H/HLA-A*02:01 mixture was evaluated by size exclusion chromatography on a Superdex™ 200 Increase 10/300 column (GE Healthcare, 28990944). The fractions of- 98% pure pHLA-A*02:01-H2-Fab complex were pooled, concentrated to 12.6 mg/mL and exchanged into a buffer containing 25 mM HEPES, pH 7.0, 200 mM NaCl.

Crystallization, data collection and structure determination

Crystals of the ternary complex H2-Fab-p53^R175H/HLA-A*02:01 were grown by vapor diffusion in hanging drops set up with a TTP mosquito robot with a reservoir solution of 0.2 M ammonium chloride and 20% (w/v) PEG 3350 MME. Crystals were flash-cooled in mother liquor. Data were collected at National Synchrotron Light Source-II at beamlines 17- ID-l(AMX) on a Dectris EIGER X 16M detector. The dataset was indexed, integrated and scaled using fastdp, XDS, and aimless. Monoclinic crystals of H2-Fab-p53^R175H/HLA- A*02:01 diffracted to 3.5 A. The structure for the H2-Fab-p53^R175H/HLA-A*02:01 complex was determined by molecular replacement with PHASER using PDB ID 604Y and 6UJ9 as the search models. The data were refined to a final resolution of 3.5 A using iterative rounds of refinement with REFMAC5 and manual rebuilding in Coot. Structures were validated using Coot and PDB Deposition tools. The model has 95.2% of the residues in preferred and 3.8% in allowed regions according to Ramachandran statistics (Table 20). Figures were rendered in PyMOL (v2.2.3, Schrodinger, LLC, New York, NY). Buried areas were calculated with PDBePISA. The docking angle that determines the relative orientation between the pHLA and the Fab/TCR was calculated by the web server TCR3d.

Mouse xenograft model

Female NOD . ( -I rkdc^{sc l}II2rg^{lin 1 ,l/V}/SzJ (NSG) mice at 6-10 weeks were acquired from the Jackson Laboratory (Bar Harbor, Maine, 005557) and treated in compliance with the institutional Animal Care and Use Committee approved protocol. In the early treatment model, mice were inoculated intravenously with 1 x 10⁶ luciferase-expressing KMS26 or KMS26-ZP53 KO cells and 1 x 10⁷ in vitro expanded human T cells via lateral tail vein injection on day 0. On day 1, mice were randomized based on luminescence quantification using the IVIS imaging system and Living Image software (Perkin Elmer) to ensure similar pretreatment tumor burden. Prior to imaging, mice received intraperitoneal injection of luciferin (150 pi, RediJect D-Luciferin Ultra Bioluminescent Substrate, PerkinElmer,

770505) were anesthetized using inhaled isoflurane in an induction chamber for 5 minutes. After randomization, two-week micro-osmotic pumps (ALZET, Cupertino, CA, 1002) filled with H2-scDb, isotype control scDb (scFv against an irrelevant pHLA linked with UCHT1 scFv), or vehicle only that had been primed in 1 mL PBS overnight at 37°C were placed intraperitoneally using sterile surgical technique. Tumor growth was serially monitored by bioluminescent imaging. In the established tumor model, mice were inoculated with 3.5 x 10⁵ or 5 x 10⁵ luciferase-expressing KMS26 cells and 1 x 10⁷ human T cells via lateral tail vein injection on day 0. On day 6, H2-scDb or isotype control scDb was administered similarly as in the early treatment model.

For mouse blood-based analysis, 200 pL blood was collected in EDTA-treated microvettes (Sarstedt, Niimbrecht, Germany, 20.1278.100) by cheek bleed, followed by centrifugation at 1000 x g for 3 minutes. Plasma was collected and stored at -80 °C until analysis. The blood cell pellet was resuspended with 100 pL PBS, followed by two 5-minute incubations with 1 mL ACK lysis buffer (Thermo Fisher Scientific, A1049201) with one PBS wash in between, and resuspended in flow stain buffer with TruStain FcX (anti-mouse CD16/32) antibody (BioLegend, 101320) and cell-surface staining antibodies. For scDb quantification, plasma was thawed and incubated in biotinylated recombinant human CD3e/5 coated streptavidin plate and detected as described in “ELISA.”

Statistical analysis

Data are presented as means ± SD unless otherwise specified. Statistical analyses were carried out using specific tests indicated in the figure legends. A P value of < 0.05 was used to denote statistical significance. All analyses were performed using Prism version 8.0 (GraphPad, San Diego, CA). In all figures, NS, P > 0.05; * P < 0.05; ** P < 0.01, ***P <

0_.001, **** p < 0_.0001_.

Example 5: Exemplary bispecific molecules targeting a TP53 mutation Table 23. Amino acid sequences of bispecific molecules targeting a TP53 mutation

OTHER EMBODIMENTS

It is to be understood that while the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A molecule comprising an antigen-binding domain that can bind to a peptide-HLA complex, wherein said peptide is derived from a modified p53 polypeptide.

2. The molecule of claim 1, wherein said modified p53 polypeptide comprises from 7 amino acids to 25 amino acids.

3. The molecule of claim 2, wherein said modified p53 polypeptide comprises 9 amino acids.

4. The molecule of any one of claim 1 to claim 3, wherein said modified p53 polypeptide comprises an amino acid sequence set forth in SEQ ID NO:l.

5. The molecule of claim 4, wherein said antigen binding domain comprises an amino acid sequence set forth in SEQ ID NO: 137, SEQ ID NO: 138, SEQ ID NO: 139, SEQ ID NO: 140, or SEQ ID NO: 141.

6. The molecule of any one of claim 1 to claim 5, wherein said molecule is selected from the group consisting of an antibody, an antibody fragment, a single chain variable fragment (scFv), a chimeric antigen receptor (CAR), a T cell receptor (TCR), a TCR mimic, a tandem scFv, a bispecific T cell engager, a diabody, a single-chain diabody (scDb), an scFv- Fc, a bispecific antibody, and a dual-affinity re-targeting antibody (DART).

7. The molecule of any one of claim 1 to claim 6, wherein said molecule further comprises an antigen-binding domain that can bind to an effector cell receptor selected from the group consisting of CD3, CD28, CD4, CD8, CD16a, NKG2D, PD-1, CTLA-4, 4-1BB, 0X40, ICOS, and CD27.

8. The molecule of claim 7, wherein said antigen-binding domain that can bind to an effector cell can bind to CD3, wherein said antigen-binding domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 170, SEQ ID NO: 171, SEQ ID NO: 172, SEQ ID NO: 173, SEQ ID NO: 174, SEQ ID NO: 175, SEQ ID NO: 176, SEQ ID NO: 177, SEQ ID NO: 178, SEQ ID NO: 179, SEQ ID NO: 180, SEQ ID NO: 181, SEQ ID NO: 182, and SEQ ID NO: 183.

9. A molecule comprising an antigen-binding domain that can bind to a peptide-HLA complex, wherein said peptide is derived from a modified RAS polypeptide.

10. The molecule of claim 9, wherein said modified RAS peptide comprises from 7 amino acids to 25 amino acids.

11. The molecule of claim 10, wherein said modified RAS peptide comprises 10 amino acids.

12. The molecule of any one of claim 9 to claim 11, wherein said modified RAS peptide comprises an amino acid sequence set forth in SEQ ID NO:2, SEQ ID NO:3, or SEQ ID NO:4.

13. The molecule of claim 12, wherein said modified RAS peptide comprises SEQ ID NO:2, and wherein said antigen binding domain comprises an amino acid sequence set forth in SEQ ID NO: 142, SEQ ID NO: 143, SEQ ID NO: 144, SEQ ID NO: 145, SEQ ID NO: 146, SEQ ID NO: 147, SEQ ID NO: 148, or SEQ ID NO: 149.

14. The molecule of claim 12, wherein said modified RAS peptide comprises SEQ ID NO:3, and wherein said antigen binding domain comprises an amino acid sequence set forth in SEQ ID NO: 150, SEQ ID NO:151, SEQ ID NO:152, SEQ ID NO:153, SEQ ID NO:154, SEQ ID NO: 155, SEQ ID NO: 156, SEQ ID NO: 157, SEQ ID NO: 158, SEQ ID NO: 159, or SEQ ID NO: 160.

15. The molecule of claim 12, wherein said modified RAS peptide comprises SEQ ID NO:4, and wherein said antigen binding domain comprises an amino acid sequence set forth in SEQ ID NO: 161, SEQ ID NO: 162, SEQ ID NO: 163, SEQ ID NO: 164, SEQ ID NO: 165, SEQ ID NO: 166, SEQ ID NO: 167, SEQ ID NO: 168, or SEQ ID NO: 169.

16. The molecule of any one of claim 9 to claim 15, wherein said molecule is selected from the group consisting of an antibody, an antibody fragment, a single chain variable fragment (scFv), a chimeric antigen receptor (CAR), a T cell receptor (TCR), a TCR mimic, a tandem scFv, a bispecific T cell engager, a diabody, a single-chain diabody (scDb), an scFv- Fc, a bispecific antibody, and a dual-affinity re-targeting antibody (DART).

17. The molecule of any one of claim 9 to claim 16, wherein said molecule further comprises an antigen-binding domain that can bind to an effector cell receptor selected from the group consisting of CD3, CD28, CD4, CD8, CD16a, NKG2D, PD-1, CTLA-4, 4-1BB, 0X40, ICOS, and CD27.

18. The molecule of claim 17, wherein said antigen-binding domain that can bind to an effector cell can bind to CD3, wherein said antigen-binding domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 170, SEQ ID NO: 171, SEQ ID NO: 172, SEQ ID NO: 173, SEQ ID NO: 174, SEQ ID NO: 175, SEQ ID NO: 176, SEQ ID NO: 177, SEQ ID NO: 178, SEQ ID NO: 179, SEQ ID NO: 180, SEQ ID NO: 181, SEQ ID NO: 182, and SEQ ID NO: 183.

19. A method for treating a mammal having a cancer, said method comprising: administering to said mammal a molecule of any one of claim 1 to claim 20, wherein said cancer comprises cancer cells expressing said modified peptide.

20. The method of claim 19, wherein said mammal is a human.

21. The method of any one of claim 19 to claim 20, wherein said cancer is selected from the group consisting of Hodgkin’s lymphoma, non-Hodgkin’s lymphoma, acute myeloid leukemia, acute lymphoblastic leukemia, multiple myeloma, a myelodysplastic syndrome (MDS), a myeloproliferative disease, lung cancer, pancreatic cancer, gastric cancer, colorectal cancer, ovarian cancer, endometrial cancer, biliary tract cancer, liver cancer, breast cancer, prostate cancer, esophageal cancer, stomach cancer, kidney cancer, bone cancer, soft tissue cancer, head and neck cancer, glioblastoma multiforme, astrocytoma, thyroid cancer, germ cell tumor, and melanoma.